University of Göteborg

November 1995






Christiane Pankow (ed.)


Dept of Linguistics

Dept of German and Dutch







SSKKII Report 9604

logo.gif (7878 bytes)

SSKKII/Dept of Linguistics/Dept of German and Dutch

University of Göteborg

BOX 720

S-405 30 Göteborg

ISBN 91-630-4786-1






On Proximity and Distality as Mood Determiners in German Reported Speech


Richard HIRSCH

The Act of Speaking: Spoken Language and Gesture in the Determination of Definiteness of Intention


Helmut KLUGE

A Chain Hanging in the Margin



Left and Right in the Perception of Picture Sequences



Shifting Worlds or Deictic Signs in www




The Semiotics of Swedish Sign Language


Christiane PANKOW

The Deictic Function of the German Adverb SO. Some Remarks on one Chapter in Karl Bühler’s Theory of Language



Humans as Signs: Iconic and Indexical



Considerations on Gombrich’s Meditations




Indexicality as Perceptual Mediation





When referring to an object, a sign can be an icon, index or symbol. Ch. S. Peirce calls this trichonomy the most important classification of signs. In modern semiotics the sign - function relation (or the sign - object relation) has become a crucial issue.

The index seems to be the most complex type among these three types of signs. Indices are the indicating signs. Indication is the simplest and basic type of semiosis. But indexical signs also play a role in very complex sign processes, such as oral and written verbal communication.

In November 1995, the Swedish Society of Semiotics (sffs) invited a small group of Swedish researchers from different disciplines and fields of research to the symposium "Indexikala tecken" [Indexicality] held at the University of Göteborg. Roland Posner (Research Center for Semiotics, Technical University of Berlin) red the introducing paper (Humans as signs: iconic and indexical). The following papers were given in order of appearance:

Shifting worlds. Deictic signs in www (Bilyana Martinovski, Göteborg), On proximity and distality as mood determiners in German reported speech (Sven-Gunnar Andersson), The deictic function of the German adverb SO. Some remarks on one chapter in Karl Bühler’s Theory of language (Christiane Pankow, Göteborg), The act of speaking: spoken language and gesture in the determination of depth of intention (Richard Hirsch, Linköping), Indexicality in the movies: space, time, and plot (Marlene Landsch, Berlin), The quotation mark and its margin line index (Helmut Kluge, Göteborg), The indexical ground of discursive reference (Tom Andersson, Uppsala), The semiotics of the Swedish sign language (Kerstin Nelfelt, Göteborg), Indexicality and material fiction (Göran Rossholm, Stockholm), Indexicality as perceptional meditation (Göran Sonesson, Lund).

In this volume selected articles based on the symposium papers are included.


Göteborg, August 1996.

Christiane Pankow



Sven-Gunnar Andersson

University of Göteborg






1. Some basic assumptions

A semiotic approach is necessary to deal with the way the proximity/ distality distinction is involved in determining mood in German reported speech. A typical characteristic of reported speech is a set of different grammatical categories and semanto-pragmatic aspects which are kept together by being linked to a dichotomy of proximity/distality with regard to the "I-here-now" of the person reporting. This is clearly deixis and thus part of indexicality, which justifies a discussion within the framework of this symposium.

The set contains such disparate elements as time reference and tense, persons involved, kind of proposition, attitudinal preferences of the person reporting, situational setting extending from private oral to official written communication.

Using one of two possible grammatical moods (indicative or subjunctive) for reported speech in German, the person reporting expresses different attitudes towards the content of the speech reported. The indicative is used in connection with an attitude of proximity on the part of the person reporting, whereas the subjunctive expresses an attitude of non-involvement, i.e. distality. There is a tendency to use the indicative more when the factors signalize proximity whether it be objectively or subjectively (cf.below). Conversely, if there are more distality factors, the greater the tendency to use the subjunctive. In cases where both kinds of factors are present, the choice of mood can be regarded as the result of a competition, where the collective weight of one kind of factor out-weighs the collective weight of those of the other kind. Within certain, as yet not too well-defined limits, the person reporting can subjectively assign special weight to some factors, thus making the balance turn over in favour of one mood rather than the other. On the other hand, there are some well-defined situations, e.g. journalists reporting public affairs in newspapers, where the use of the subjunctive is conventionally fixed in a way which allows practically no exceptions. This is determined by a pragmatic principle: the reporting journalist is in no way to interfere with the content of what is being reported.

2. Some basic examples

The following is in part an adaptation and elaboration of an example given by Engel (1988, 110-11). Supposing that in a university department two people, one a professor, the other a teaching assistant, are discussing how their subject is taught at the neighbouring university, the professor may criticize the quality of the teaching, saying:

(1) "Dort kann man nicht viel lernen"

[There can one not very much learn]

On the next day the teaching assistant is talking to a student, who is planning to move to the neighbouring university. During their conversation the teaching assistant says:

(2) Gestern sagte mir aber einer,

[yesterday said to me, though, somebody]

daß man dort nicht viel lernen könne / kann / konnte(?) / könnte

[that one there not very much learn can / can / could / could]

What he does by saying (2), at least when using the present subjunctive, is rendering the content of somebody else's utterance in such a way that it is fully clear that he is doing just that. He refers to a former act of communication by Gestern sagte , to a person as having uttered something (sagte ... einer ) and then gives the content of the utterance in the form of a subordinate clause with the complementizer daß 1). The subordinate clause contains the information ascribed to the original speaker though not in exactly the same form as in the (real or postulated) original utterance. Quotation of direct speech would be:

(3) Gestern sagte aber einer zu mir: „Dort kann man nicht viel lernen."

[yesterday said though somebody to me (cf. (1))]

Total identity is claimed between the quotation and the original (or model) utterance in terms of lexical and grammatical structure.

To be identified as such, both reporting and quoting require at least one index. The minimum seems to be reference to a source of information other than the present speaker/author at the moment of formulating the text. If the teaching assistant had said (1) without further comments, concerning the quality of the neighbouring university, there would have been no index of rendering and thus no hint whatsoever that he was quoting. Furthermore, the proposition would have been asserted by himself, thus letting the student know that he was of that negative opinion about the neighbouring university. Even if there is a possibility of interpreting (1), when uttered by the teaching assistant, as the description of a well-known fact or as a wide-spread opinion, the latter, by uttering (1), declares himself as one who shares this opinion as long as he does not add anything to modify his utterance.

Returning to examples (2) and (3) there is an important difference to be noticed, which is of concern for this article. In German reported speech there is a potential variability of tense and mood of the finite verb when reporting, whereas in quotation, of course, the verb form of the original utterance has to be retained. The variability of tense and mood in (2) is not a case of free variation. The differences between the forms correlate - it is contended here - with differences in the proximity/ distality dimension. This will be discussed below.

3. Discussion

3.1 The present subjunctive

Using the present subjunctive (könne) in (2), the person reporting overtly and unmistakably declares that the content of the clause is to be taken as something uttered by someone else. The present tense subjunctive thus marks the information as depending exclusively on the original speaker. This is clearly a distality function, dissociating the information from the person reporting. It follows that the latter takes no responsibility for its truth value and does not take a stand on the issue about its being justified or not. At least this is what is said by the theory and in handbooks. It certainly fits in very well with the usage in serious journalism and official protocols of different kinds, where there is a situationally and conventionally established distance of non-involvement between the reporter and the communication reported.

In our example, however, the situation is different. The teaching assistant reports something which was said to him by his colleague . His colleague as a speaker and he himself as the person spoken to, constitute two proximity factors. By not giving away his colleague and using einer 'someone' instead, the degree of proximity is diminished. There is, however, another proximity factor at the moment when the teaching assistant utters (2) viz. the existence of the alleged low quality of education at the neighbouring university. Using the present subjunctive - a distality form - for reporting a proposition which describes a situation which must be taken as simultaneous with the moment when the teaching assistant utters (2) - simultaneity of this kind being a proximity factor - conveys the impression that he is distancing himself from the information reported, i.e. from its truth value. Everything else calls for proximity: the teaching assistant saying something he himself was told in the original act of communication, the fact that (2) is uttered in a face-to-face communication between two people interested in the subject matter talked about, the situation at the neighbouring university being simultaneous with the moment of uttering (2), the likelihood of the teaching assistant knowing the education quality of his own subject at the neighbouring university. There only remains one possible reason for him to use the distality form könne : to mark distality by ascribing the information totally to its source. At the same time, this gives the subjunctive form such weight that he seems to choose sides on the truth issue by using, paradoxically, a form whose systemic meaning is viewed as a signal of just reporting objectively without choosing sides. Not choosing sides, can however, also be understood as a kind of choice - like distancing oneself.

3.2 The present indicative

The present indicative kann in

(2I) Gestern sagte mir aber einer, daß man dort nicht viel lernen kann

allows for at least three different interpretations: 1. The teaching assistant accepts the professor's assertion and thus stands behind it too; 2. The teaching assistant refers to the simultaneity between the clause content and his own moment of uttering (2); 3. The teaching assistant just takes over the tense and mood of the original utterance (1), creating a type of rendering which is intermediate between reported speech and quotation.


3.2.1 Taking over the proposition

Under the interpretation of taking over the professor's assertion, the verbum dicendi of the main clause (sagte (mir)) does not merely mean 'uttered to me', but also 'told me, informed me'. Moreover, the teaching assistant gives the proposition "man kann dort nicht viel lernen" ('you can't learn very much there') its full truth value which means that he does not refer it only to its source, that is to the professor. An adequate paraphrase of

(2I) under this reading is

(4) Dort kann man aber nicht viel lernen. Das sagte

There can one though not very much learn. That said

mir gestern einer.

to me yesterday someone

Under the reading of (2I) as paraphrased by (4) the teaching assistant shortens the distance between the professor's utterance and his own. The use of "shortens" means that he does not mark it as exclusively the text of someone else but brings the statement into his own set of opinions, thus allowing it to go for his own, too. The subjunctive könne (cf. 3.1.) points unambiguously into the world of opinions of the original speaker (in this case the professor) but the indicative kann does not. Engel (1988, 115) says that through the indicative the secondary text - the speech reported - is at least in part lifted up to the level of the primary text and thereby counts as the text of the reporting speaker, too. The latter, so Engel says, partially identifies himself with the secondary text. In any case the speaker reporting (in our case the teaching assistant) must reckon with an attentive listener to understand his communication in this way.

The interpretation of (2I) as represented by (4) can, in my view, be taken to mean that the indicative morpheme of the present indicative form kann as opposed to the present subjunctive könne has the proximity value of also referring the proposition reported to the person reporting by marking it as part of his set of opinions. If the "I-here-now" of the speaker is taken as the basis of deixis, then this represents the case of reference to the speaker's "I" by the present indicative.


3.2.2 Simultaneity with the moment of speech

The present indicative also denotes that something is the case at the moment of speech. This is quite frequently taken to be the basic systemic meaning of the present tense. The question whether that is really the case will be left out of this discussion, which focusses on the content of (2I). The speech reported there contains a general statement about the quality of education at the neighbouring university. This means that the deplorable state of affairs as described by the professor in (1) and reported in (2I) is valid for some period of time. We know from our knowledge of the world and from our knowledge of the rules of communication (as for example stated by Grice) that the period of validity extends from yesterday when the professor uttered (1) over the following day when the teaching assistant uttered (2I) up to and inclusive of the presumed future time to be spent by the student at the neighbouring university. Mentioning the matter would be fairly pointless otherwise.

Under this reading the present indicative kann in (2I) refers from a strictly time-deixis point of view to the "now" of the speaker reporting (i.e. the teaching assistant), regardless of his attitude towards its truth value. The proximity value of the present indicative here is proximity with regard to the "now" of the teaching assistant, that is simultaneity with his counseling conversation with the student.

3.2.3 Intermediate between report and quotation

There is one reading which is entirely different from those above, both of which have the teaching assistant as the origo of deixis. The difference consists in taking the original speaker as origo of deixis, in our case the professor. Under this reading kann in (2I) is merely a copy of the form kann in (1). Taking over a concept from the field of entomology this case might be called "incomplete transformation", incompleteness here being the keeping of tense and mood of the original utterance. This is an intermediate form between reported speech "proper" and quotation and is very common in everyday spoken German. In the following example it is the only possible interpretation:

(5) Vor zwei Wochen sagte er, daß er immer noch krank ist

Two weeks ago said he that he still ill is (pres.ind.)


The original utterance must be reconstructed as:

(6) „Ich bin immer noch krank"

I am still ill (pres.ind)

The case discussed in this section (explicated by the unambigous example (5)) illustrates proximity to the original speaker's original utterance by retaining tense and mood (as abstract categories, not as forms, cf. (5) and (6)). Pronominal deixis is shifted, if there is one to be shifted as in (5) (in (1) there is not).The reason for this is strictly pragmatic: it must be made clear who is being talked about.

Retaining tense and mood of the original utterance means proximity to the "now" of the original speaker.

3.3. The preterite indicative

In the presentation of (2) in section 2 above, a question mark was put after the preterite indicative form konnte ('could') in the clause containing the reported speech, cf. (2II):

(2II) Gestern sagte mir aber einer, daß man dort nicht

Yesterday told me, though, someone that one there not

viel lernen konnte (?)

much learn could (pret.ind.)

The question mark is due to the fact that there is very much doubt among native speakers as to whether this really renders the original utterance:

(1) „Dort kann man nicht viel lernen"

There can one not much learn (pres.ind.)

The normal interpretation of (2II) takes as its starting point:

(1I) „Dort konnte man nicht viel lernen" (could)

which leaves us with the case of copying the tense/mood of the original utterance as just described in section 3.2.3.

Konnte in (2II) as a counterpart of kann in (1) would illustrate the kind of tense adjustment normal for English and Swedish (adjustment to the tense of the superordinate clause). The distality of the preterite indicative konnte in German is purely time referential in character, not modal as in English and Swedish (which both lack a subjunctive and therefore use the distality element of the preterite for both time and modality functions). Under the reading of (1) only that segment of time would be referred to by konnte in (2II) which is simultaneous with the time of the original utterance by the professor and which is distant from the point in time when the teaching assistant utters (2II). Since, obviously, for pragmatic reasons the actuality of the proposition is of importance it would be odd to mark it as not actual, i.e. as distant in time. This is one of the reasons why (2II) for most Germans is no rendering of (1) but of (1I) which describes quite another situation, namely, someone looking back at his previous stay at the neighbouring university.

3.4. The preterite subjunctive

There is a possibility of the teaching assistant saying:

(2III) Gestern sagte mir aber einer, daß man dort

Yesterday told me, though, someone that one there

nicht viel lernen könnte

not much learn could


From a purely morphological point of view, the form könnte contains two morphemes of distality, the subjunctive element as expressed by the umlaut vowel -ö - and the morphologically preterite element -te . Depending on the speaker's variety of German, there is no difference in reported speech between the present subjunctive (könne ) and the preterite subjunctive (könnte ), the present subjunctive being merely very much less used, or there is one in cases where both forms can occur. In the latter case the preterite subjunctive conveys a higher degree of modal distality (reflecting the double morphological distality marking), the speaker distancing himself from the alleged truth value of the proposition reported, wanting to say that the assertion of the original speaker is open to doubt. It is obviously not easy for a listener to know which variety is actually being used. This has led to a situation where it takes a lot more than the bare preterite subjunctive form for reported speech to be interpreted in the "strong" way just described.

Moreover, there is another kind of distality involved, at least for the listener. The form könnte in (2III) could also be understood as reflecting a könnte in the original utterance:

(1II) Dort könnte man aber nicht viel lernen

There could one, though, not much learn


Under this reading the double distality of könnte in both (1II) and (2III) conveys counterfactual meaning: 'You would not be able to learn very much there'. Such an interpretation would, in this case, be completely justified from a communicative and pragmatic point of view. Therefore, using könnte leaves room for several distality interpretations whereas the present subjunctive könne marks reported speech only and no counterfactual meaning (cf. section 2).

4. Proximity/distality in the superordinate clause

In the type of examples discussed so far, the verb of the superordinate clause is in the preterite tense, that is in a distality form. It seems as if it takes a lot of proximity factors to outweigh this distality to use the proximity form of the present indicative instead of the subjunctive distality forms in the clause containing the reported speech. In the examples above such proximity factors are: 1st person being spoken to, the fact referred to by the proposition of the subordinate clause being plausible to the reporting speaker and being the case still at the moment of report, i.e. being simultaneous with it. If the verb of the superordinate clause is, by contrast, in the present

tense, that is in a proximity form, then the indicative is free of choice in the subordinate clause in all varieties of German (written or spoken) without the necessity of there being other supplementary proximity factors. Thus using the present for reporting, the teaching assistant might render (1) as

(2IV) Da sagt mir aber gestern einer, daß man dort nicht viel lernen kann (tells)

that is with the historical present, or as

(2V) Man sagt aber, daß man dort nicht viel lernen kann

(people say)

with the general present.

5. Concluding remarks

From the discussion in the preceding sections follows that the use of one verbal form rather than another when reporting, means expressing a certain evaluation of the proposition in terms of proximity/distality. The different verbal forms do not replace each other without consequences as to this dichotomy. Instead of talking simply of different types of reported speech one could take the person reporting as a starting point, saying that the producer of a text places information stemming from other communication acts in his own text with different degrees of integration, i.e. with different degrees of proximity/distality between his own text and what is taken over from another source.

Deixis seems to be a useful tool for the analysis of factors determining tense and mood of the text taken over: the person morpheme, the subjunctive morpheme and the indicative morpheme, the present tense morpheme and the preterite tense morpheme point to different positions and distances of the information taken over in relation to the origo of the "I-here-now" of the person using the information - at least in cases where the language system of German provides a full differentiation. At what point in this psychological space the information is situated by the speaker, that is, what verbal form is chosen is - and this is what I want to stress in this paper - a function of the proximity and distality factors of the whole utterance (superordinate + subordinate clause).

The parameter of distality also quite clearly lies behind the rule of choice for tense/mood form of the finite verb in reported speech in the neutral register of written non-fictional prose and in non-commenting newspaper reports when the superordinate clause is in the preterite tense:

1.Use the subjunctive form if it is morphologically unambiguous and then in the following order: use the present subjunctive if it is unambiguous, otherwise the preterite subjunctive.

2.If both the present and the preterite subjunctive are morphologically ambiguous, that is if they coincide with the indicative, then use the preterite.

This rule, which is very much adhered to, obviously rests on the distality of the subjunctive mood and on the distality of the preterite tense, combining the two morphological categories only through their common deictic force of distality. Quod erat demonstrandum .


1) In this article there will be no discussion of reported speech without a complementizer (as in: Gestern sagte mir aber einer, man könne / kann dort nicht viel lernen).




Andersson, S.-G. 1994a: "Proximität und Distalität im deutschen Tempus/Modussystem", in: Nordlyd (Tromsø University Working Papers on Language and Linguistics) 22, 1-7.

Andersson, S-G. 1994b: "Zum Indikativ in eingeleiteten Sätzen der indirekten Rede nach präteritalem Anführungsausdruck", in: Nordlyd 22, 38-52.

Engel,U. 1988: Deutsche Grammatik . Heidelberg (Groos).



Richard Hirsch

Linköping University







The roman rhetorical tradition acknowledged the importance of gesture and made the appropriate use of gesture an important part of the 'actio' of a speech. Quintillian devoted a large portion of one of the four books of his Institutio Oratoria to a discussion of the proper use of gesture by an orator. Mainstream modern linguistic theorizing has had a condescending or downright antagonistic attitude toward gesture. Due to a Cartesian dualistic bias where body and mind are strictly separated and to a concentration on the enterprise of accounting for linguistic competence rather than linguistic performance, gesture occurring in connection with spoken language has generally been ignored as irrelevant.

This situation is however changing. Linguists are coming together with communication scientists, anthropologists, psychologists, and others to study the actual use of spoken language in a variety of everyday situational contexts. In this regard I would refer the reader to the excellent work being done by Charles and Marjorie Goodwin on the analysis of video recordings of language use in natural settings (cf. Goodwin & Goodwin 1992). Recent psycholinguistic research shows that speech and gesture are probably neurophysiologically related (cf. McNeill 1992 and Feyereisen & de Lannoy 1991).

Most studies of the integration of gesture and speech have been 'syntactically' oriented, i.e. determining the temporal order of occurrence of the gesture and the corresponding speech segments. Usually as a effort to investigate the process of speech production in relation to thought (cf. Feyereisen & de Lannoy 1991). McNeill (1992) has however started to move in a more semantic direction and has studied the use of illustrative and metaphoric 'imagistic' gestures in connection with speech.

In contrast to Decartes, C.S. Peirce realized that knowledge or cognition has three basic semiotic dimensions; iconic, indexical, and symbolic. Peirce claimed that these three dimensions of cognition were grounded in intuitions of similarity, causality, contiguity in space-time and part-whole, and arbitrary conventional connections between objects (abstract or concrete) of attention. In a Peircian semiotics the iconic and indexical dimensions of signs are primarily non-verbal, the symbolic dimension is primarily verbal.

The question to be addressed in this paper is how the non-vocal, non-verbal aspects of gestures are related to the vocal, verbal aspects of spoken language (speech). The relationship that will be explored and discussed is a semantic one. To this end I have turned to Arne Naess's (1953) Theory of Interpretation and Preciseness for inspiration. Naess's theory is meant to be a tool for semantic analysis of communicative language use either spoken or written. I am going to generalize Naess's semantic insights in a semiotic direction to cover both gesture and speech.

Terminology and Definitions

Naess (1953) uses the term expression to refer to a linguistic formulation, usually a statement or a sentence. I use expression in a wider sense to refer to a linguistic vocal, verbal sign and/or a non-vocal, non-verbal sign. Based on a reformulation of Grice's (1957) distinctions between natural and non-natural meaning to be found in Allwood (1976) and Hirsch (1989) I define expression as follows.


Definition: A noteable movement of the body or part's of the

body that,

i) indicates, i.e. functions as a source of information to an observer

ii) displays, i.e. is intended to make a receiver at least apprehend or attend to certain information, through some manner of apprehension like direct perception or inference.

iii) signals, i.e. is intended to make a receiver at least apprehend a display of certain information, through direct perception or inference. (cf. Grice's notion of Meaning NN )

iv) symbolizes, i.e. functions by convention as a representative displayor of information.


Different types of expression are related to the distinctions verbal, non-verbal, and vocal, non-vocal as illustrated in table 1.

Table 1: Expression Types

Vocal Non-vocal


Verbal Spoken Written

Language Language


Non-verbal Vocalizations Gestures

As can be seen in table 1, gestures are non-verbal, non-vocal expressions. I define gesture for the purposes of this paper, as follows.


Definition: semiotic phenomena characterizable as non-conventional, non-vocal, nonverbal, non-alter contact communicative behavior produced by movements and/or configurations of the upper extremities of the body, i.e. the hands, fingers, arms, shoulders and head.

The aim of analysis in a Naessian semantics is to gain insight into how people go about dealing with communication problems involving the achievement of cognitive understanding or the avoidance of cognitive misunderstanding, i.e. how people arrive at mutually confirmed agreement about what is meant by what is expressed even though parties may disagree as to whether what is meant is true or false, practical or impractical, etc. Naess introduces the notion of definiteness of intention (which he and others following him later called depth of intention) to capture the insight that speakers can vary as to the number of cognitive distinctions that they are aware of and prepared to take into consideration in a particular situation in connection with the use of a particular expression. This holds also for the same speaker in different communicative situations.

Situational and personal relativity in the cognitive understanding connected with expressions in communication is part of a more general and fundamental insight that I sometimes refer to as the Naess semantic uncertainty principle which can be formulated in a weak or a strong version.

Weak version:

It is not necessarily the case that speakers have a particularly well determined definiteness of intention in mind when expressing themselves.

Strong version:

In many/most cases in everyday life, speakers do not have a particularly well determined definiteness of intention in mind when expressing themselves.

The semantic uncertainty principle leads to the following consequence.


What definiteness of intention speakers have/had in mind when expressing themselves can, in certain (weak)/most (strong) cases, only be determined by a process of interpretation.

Naess's semantic theory was primarily developed to deal with the interpretation of verbal - verbal expressions or formulations. For Naess, interpretation consists basically of the determination of definiteness of intention by means of various types of reformulations. My extension of Naess's theory attempts to incorporate the interpretation of verbal and non-verbal expressions within the same basic framework where the determination of definiteness of intention is effected by a combination of vocal, verbal and non-vocal, non-verbal means.

The framework I am going to use for analysis of the determination of definiteness of intention in face-to-face spoken interaction consists of the following basic assumptions and definitions. In these definitions it should be born in mind that the term expression covers both vocal, verbal and non-vocal, non-verbal expressions, i.e. speech and gesture.

1. The definiteness of intention of expressions varies according to situation and personal interpretations of the expressions.

Comment: This is a direct consequence of the Naess semantic uncertainty principle discussed above.

2. An expression T is an interpretation of an expression U if and only if there is at least one situation in which for at least one person T and U are understood cognitively in the same way or have the same cognitive effects, i.e. are equivalent in cognitive adequacy.

Comment: For instance, let T be the vocal, verbal expression "OK" and U be the gesture 'thumb-index ring'. In certain situations T:"OK" is used as an interpretation of U: 'thumb-index ring'. Especially in situations where U: 'thumb-index ring' occurs first in face-to-face spoken interaction.

3. Two expressions (T and U) are cognitively non-equivalent for at least one person in at least one situation if and only if there is one imaginable set of circumstances for which the person(s) would claim that the one expression (T) was cognitively adequate and the other expression (U) cognitively inadequate.

Comment: This can be illustrated in the case of two vocal, verbal expressions, say T:"block" and U:"cube". These two expressions are cognitively non-equivalent in a situation in which speakers are talking about a collection of blocks where some blocks are cubes and others are not. T: "block" would, for instance, be cognitively adequate as an expression referring to a pyramid, whereas U: "cube" would not. In a situation where a speaker held the gesture 'thumb-ring index' to be an expression referring to money, as for instance in Japan, T:"OK"(or the Japanese translation equivalent) and U:'thumb-index ring' would not be cognitively equivalent.

4. A necessary but not sufficient condition for equivalence of cognitive understanding is that two expressions are not cognitively non-equivalent.

Comment: In the case of vocal, verbal expressions this condition rules out T:"Good food is not cheap" being equivalent to U:"Cheap food is not good" although both can be reduced logically to the same expression, namely 'No food is both cheap and good', i.e. both can be shown to have the same truth-conditions. The expressions are, however, obviously not cognitively equivalent.

5. Two expressions (T and U) are equivalent in definiteness of intention if and only if for at least one person in at least one situation all interpretations of T are interpretations of U and all interpretations of U are interpretations of T. All ways of understanding T are going to be ways of understanding U and vice versa.

Comment: In our example T:"OK" and U:'thumb-index ring', it is possible that for at least one person in at least one situation all ways of understanding T:"OK" will be ways of understanding U:'thumb-index ring', namely 'good', 'right', 'agreed', etc.

6. Two persons have the same definiteness of intention in relation to an expression if and only if the expression is equivalent in cognitive adequacy for both persons, i.e. all ways of understanding the expression by one person are going to be ways of understanding the expression by the other person, and vice versa, in at least one situation.

Comment: This definition covers the case when we have mutual understanding and confirmed agreement on the meaning of an expression by two persons in a communicative situation.

7. Two persons have sufficiently equivalent definiteness of intention in relation to an expression if and only if they can use a set of interpretations of the expression to arrive at a mutually confirmed agreement as to the meaning of the expression in order to solve at least one communication problem in a least one situation.

Comment: This definition covers the normal state of affairs in what is usually referred to as 'successful communication' in face-to-face spoken interaction. For instance, in a certain interaction involving reference to blocks, two persons may be able to use a set of interpretations of an expression such as "the square one" to solve a communication problem in the ongoing interaction, although the concept of a square block may be conceived of as a contradiction in terms outside of the ongoing interaction.

8. An expression T is more precise or effects a greater definiteness of intention than an expression U if and only if all interpretations of T are interpretations of U and there is at least one interpretation of T which is more clearly decideable in its application and non-application to any given entity or phenomenon than other interpretations of U.

Comment: The set of interpretations of T is a proper subset of the set of interpretations of U, and the certainty of application and non-application associated with T is greater than that associated with U. All ways of understanding T:"a three dimensional closed figure with only square faces" are going to be ways of understanding U:"block" and at least one interpretation of T, e.g. "cube", is more clearly decideable than any other interpretation of U. All ways of understanding the expression T:'thumb-index ring' are going to be ways of understanding the expression U:'configuration of thumb and index finger' and at least one interpretation of T:'thumb-index ring', e.g. "agreed" is more clearly decideable than other interpretations of U:'configuration of thumb and index finger'.

9. An expression T is less precise, i.e. more vague (expresses less definiteness of intention) than an expression U if and only if all interpretations of T are interpretations of U and there is at least one interpretation of T which is less clearly decideable in its application and non-application to any given entity or phenomenon than other interpretations of U.

Comment: The set of interpretations of T is a superset of the set of interpretations of U. And the certainty of application and non-application associated with T is less than that associated with U. This is the exact inverse of the precision relation between expressions.

10. An expression T is an elaboration of an expression U if and only if all interpretations of T are interpretations of U and all interpretations of U are interpretations of T , and T is neither more nor less precise than U.

Comment: Elaborations of an expression are either some sort of paraphrases or have the effect of making the expression more or less specific rather than more or less precise (cf. Naess 1953, Pinkal 1986, Hirsch 1989).

In the next section of this paper, I am going to demonstrate with empirical evidence how speakers' vocal, verbal and non-vocal, non-verbal expressions collaborate in the incremental determination of definiteness of intention in ongoing spoken interaction. Speakers' expressions will be seen to be part of an historical process of determination of definiteness of intention within an interactional context where there is an incremental development of information involving an integration of indicative, display, signal, and symbolic expressions over the course of one speaker's turn at talk or across turns and speakers in the course of the spoken interaction.

Empirical investigation

The experimental situation consisted of a studio video recording of three students participating in a problem-solving task involving the construction of a marble track according to a specification given by me, the experiment leader. The communicative behavior exhibited by the participants was not elicited or directed by the experiment leader. The construction task involves many complex geometrical and spatial relations and can therefore be expected to generate gestures in connection with speech in the interaction (cf. Feyereisen & de Lannoy 1991). The construction specification consisted of asking the students to build a track with a set of blocks that would allow a small steel marble to roll west, south, east, and finally north so that the marble ended up directly below the position at which it started as illustrated in figure 1.

Figure 1.


In this task situation most gesturing occurs in connection with a direct manipulation of the construction blocks. I define a direct manipulation gesture as follows.


Direct Manipulation

Definition: Grasping and/or moving a block or a part of the blocks construction with one or both hands.

Table 2 contains a nonexchaustive collection of examples of cases of vocal, verbal expressions used in conjunction with non-vocal, non-verbal direct manipulation gestures which occurred in the video recorded problem-solving session. These gestures can be seen as illustrations of what Searle has called the limiting case of saying, namely saying that involves showing (Searle 1969: 88).

The direct manipulation gesture may be characterized as an expression which has less indefiniteness of intention than the preceding vocal, verbal expression. Here the speaker is actually manifesting an intention in practical action in full view of the co-participants. The direct manipulation gestures in table 2 can be classified as signal expressions according to our typology of expressions in section 1. The vocal, verbal symbolic expressions are complemented by non-vocal, non-verbal signal expressions to accomplish a demonstration of the speakers' intentions.

Table 2.

1. så <gesture> <DM> 2. så där <gesture> <DM>

like this like that

3. vi ska ha en sån <gesture> <DM> 4. ta en sån där <gesture><DM>

we need one like this take one like that

5. det måste var så <gesture> <DM> 6. den här måste också ha ett hål så

it should be like this this one has to have a hole in it so

kulan kan trilla ner <gesture><DM>

the marble can roll down

7. vi ska ha en sån <gesture> <DM> 8. ta en sån här med hål i <gesture><DM>

we should have one of these take one with a hole in it

9. sätt på en sån här istället <gesture><DM

put one like this instead 10. kan man inte byta plats <gesture><DM>

couldn't you change places

11. kan man inte sätta den under 12. om vi vänder den och sätter den så

can't you put that under if we turn it and put it like this

<gesture> <DM> istället <gesture><DM> instead

13. vi kan prova med det här 14. men om man lägger den

we can try with this but if you lay that

<gesture> <DM> <gesture> <DM>

15. den här lutar kan man säga 16. den där så <gesture> <DM>

this one slants you might say that one like this

<gesture> <DM>

17. vi skulle ha den <gesture><DM> 18. den här då kanske kan åka in i den

we should have that one this one could maybe roll into that one

<gesture> <DM>

(See note 1 at the end of the paper for transcription conventions.)

The shorter combinations of vocal, verbal and non-vocal, non-verbal expressions found in table 2 are complemented by longer stretches of intrasubjective incremental development as in example 1.

Example 1.

Frank: men <vad håller vi på med?> <torso lean back

but what are we doing? hands out palms up>

de här ä mycket smidigare att

this is much more elegant to

göra så här

do like this

<ta den> (...) <DM>

take this

<så sätter vi upp den> <DM>

and we put this up

<då gör vi så> <DM>

then we do like this

In example 1, Frank's combined vocal, verbal and non-vocal, non-verbal expression

men <vad håller vi på med?> <torso lean back

but what are we doing? hands out palms up>

can be interpreted as follows. The indicating expression <torso lean back hands out palms up> could be glossed as meaning 'here it is' or at least as some sort of offering gesture. In this instance, we have a basically indefinite display expression that is determined by integration with a vocal, verbal symbolic expression "what are we doing?". A variant of this gesture is presented by Frank, this time without accompanying vocal verbal expressions and with a different communicative function, after the appropriate adjustments to the construction have been made and the construction passes the marble test, i.e. the marble rolls around in the track according to the construction specification.

Frank: <gesture> <both hands out palms up>

The vocal, verbal question "what are we doing?" that acccompanies the expression <torso lean back hands out palms up> incorporates the non-vocal, non-verbal display into the ongoing interaction as a stretch of behavior relevant to the purpose of the problem-solving exercise.

The course of development continues with a series of direct manipulation gestures introduced and accompanied by vocal, verbal expressions.

de här ä mycket smidigare att

this is much more elegant to

göra så här

do like this

<ta den> (...) <DM>

take this

<så sätter vi upp den> <DM>

and we put this up

<då gör vi så> <DM>

then we do like this

What I want to propose is that direct manipulation gestures like those exemplified above can be characterized as non-vocal, nonverbal precisification operations on the definiteness of intention of the vocal, verbal expressions. These gestures are simply more restricted in their interpretative potential than the preceding speech expressions. There are fewer ways of understanding or interpreting these gestures than there are ways of understanding or interpreting the speech expressions. The sets of interpretations of the demonstrative direct manipulation gestures are proper subsets of the sets of interpretations of the speech expressions and there is more certainty concerning their application or non-application to the situation at hand in the ongoing interaction.

This does not mean, of course, that all such direct manipulation gestures can be viewed as precisification operations. Neither should it be taken to mean that all vocal, verbal expressions have non-vocal, non-verbal counterparts of equivalent cognitive content and vice versa. However, in this particular type of situation many vocal, verbal expressions are made more determinate as to their definiteness of intention by being combined with a direct manipulation gesture.

In example 2, we find the participants cooperating to produce what I call an intersubjective course of development of definiteness of intention. The speakers work on each others' expressions trying to better determine an initially indeterminate definiteness of intention.

Example 2

Jane: det måste vara fyra våningar högt

it has to be four stories high

<gesture> eller någonting <right index rise>

or something

Frank: för man<tappar ju en för varje><right hand illustrates because you lose one for each stepwise drop>

Linda: <går ner till den nivå sen går ner

go down to that level then go down

till den nivå> <right index illustrates

to that level stepwise drop>

The development takes place in a basically recursive manner where an expression by one speaker is made more precise, more vague, or elaborated on in some manner, either by the use of expressions from the same speaker or by expressions from other speakers. Figure 2 contains an analysis of the recursive structure of the development. The expressions labeled U are expressions that are determined in some manner or other by expressions labeled T (see the definitions above). An expression labeled T may itself consist of expressions which consist of a determined part (U) and a determiner part (T).

As figure 2 illustrates, vocal, verbal and/or non-vocal, non-verbal expressions by a speaker are subjected to determination by means of other vocal, verbal and/or non-vocal, non-verbal expressions, either by the same or another speaker. The determination is seen to be recursive in nature where, for instance, a precisification may be elaborated on or an elaboration may in turn be elaborated on.


The discussion of these examples of the integration of speech and gesture in face-to-face spoken interaction demonstrate a point made by Grice in the closing line of his 1957 article 'Meaning' where he claims that "to show that the criteria for judging linguistic intentions are very like the criteria for judging nonlinguistic intentions is to show that linguistic intentions are very like nonlinguistic intentions". Both linguistic and nonlinguistic intentions are matters of interpretation and vary in definiteness: compare the direct manipulation gestures that accompany the short vocal, verbal expressions in table 2 and the vocal, verbal accompaniment of Frank's <torso lean back hands out palms out> expression in example 1. This variation in definiteness is something that must be dealt with and worked out to the satisfaction of the participants within the ongoing face-to-face interaction.

I hope to have shown that we may profitably employ a semiotically generalized version of Arne Naess's theory of Interpretation and Preciseness to account for the interpretative work that is carried out by participants in face-to-face spoken interaction. Even if I am proven wrong in the details of the analyses offered above or even in the broad theoretical outlines, I think I am still right in saying that a more empirically oriented linguistic methodology is going to run head on into the problem of the integration of speech and gesture and that there must be a place provided in linguistic theorizing for gestures in

combination with speech.


1. Transcription conventions:

Vocal, verbal expressions in the left-hand column of the transcription within < > brackets are accompanied by non-vocal, non-verbal expression to be found directly to the right within matching < > brackets.

<words in transcription> <description of non-vocal, non-verbal expression>

The cases in which non-vocal, non-verbal expressions occur without overlapping with vocal, verbal expressions is indicated as follows.

<gesture> <description of non-vocal,

non-verbal expression>

Inaudiable speech is indicated by dots enclosed in parentheses as follows.






Goodwin, C. & Goodwin, M.H. (1992) 'Context, activity, and participation'. in P. Auer and A. di Luzio (eds.) The Contextualization of Language. Amsterdam: John Benjamins.

McNeill, D. (1992) Hand and Mind What Gestures Reveal about Thought. Chicago: University of Chicago Press.

Feyereisen, P. & de Lannoy, J-D. (1991) Gestures and speech: Psychological investigations. Cambridge: Cambridge University Press.

Searle, J. (1969) Speech Acts. Cambridge: Cambridge University Press.

Allwood, J. (1976) Linguistic Communication as Action and Cooperation. Gothenburg Monographs in Linguistics 2. Dept. of Linguistics. University of Göteborg.

Hirsch, R. (1989) Argumentation, Information, and Interaction. Gothenburg Monographs in Linguistics 7. Dept. of Linguistics. University of Göteborg.

Naess, A. (1953) Interpretation and Preciseness. Olso: Universitetsforlaget.

Peirce, C.S. (1955) 'Logic as Semiotic: The Theory of Signs' in Buchler, J. (ed.) Philosophical Writings of Peirce. New York: Dover.

Grice, H.P. (1957) 'Meaning'. Philosophical Review, 66, 377-388.

Pinkal, M. (1986) Logik und Lexikon. Berlin: de Gruyter.





Helmut Kluge

University of Göteborg






0. On the placement of quotation marks

I'd like to begin with a gesture. I won't make the gesture myself, because it is a dubious one, but I am going to describe it. I don't really mean anything by this gesture either, I just want to mention it. I have therefore two reasons not to show it.

The gesture in question is about fifty, sixty years old. That is a young age in the history of gestures. It originates from intellectual circles in the USA and is now widely used in the States as well as in Europe.

The gesture to be described was invented as a substitute for quotation marks. Because quotation marks and their development throughout many centuries will be the subject of this paper, a modern gesture substituting them is of some interest.

Before presenting this gesture I would like to remove some prejudices regarding the quotation mark and to name a theory of its origin. I do hope this will startle you and arouse your interest.

1. Two ages of quotative marking

In the avarage compendium on punctuation one finds that quotation marks are described as signs uniform throughout the centuries. Supposedly they are unequivocal signs of punctuation and they are always related to quotations. Opening up a medieval codex, however, you will see that this is not true. So let me begin with some preliminary statements on their status, their function and their placement.

The quotation mark was not a sign of punctuation before 1650. Therefore it is hardly mentioned in treatises on syntax or punctuation written earlier than the 17th century.

The quotation mark has enclosed a large variety of items

throughout history, not only quotations. Thus being named incorrectly, it has seldom been described adequately. It might be advisable to find a new name for the earlier sign.

The quotation mark before 1650 was not a pair of signs someplace in the middle of lines, but rather a linear group of signs running in the margin and in correspondence with the text proper. By no means was it the sign closely sourrounding a quoted item, like we are used to today. The old mark was written vertically.

Quotation marks probably originated in the text criticism of the late classical period of Greece. They can be traced back to the Hellenistic schools in Egypt, mainly to the pagan, to the Jewish and to the Christian schools of Alexandria.

From about 200 A.D. there is a straight forward development of quotation marks. Below I will present a list of books containing quotation marks from each century since their introduction.

The continuity is articulated in many ways. There have been a great deal of changes in their placement, shape, and function.

1.0.1 A quotative gesture

To put down quotation marks on papyrus, parchment, paper, or on the screen is one thing. To see them quivering in the air is another thing . I am sure you know what I mean, because you all have seen this gesture or even do it yourself, a dozen times every day. Imagine these speakers: just when their ordinary stream of information is transponed into quoting or joking, into supposition or irony, they suddenly raise their arms, the hands at head-level, and then something strange happens. Smiling conspiratively they start jiggling their forefingers and middlefingers. We have learnt to understand this maneuver, because of its strong pragmatic impact. What has happened? Apparently their brain went on strike, their tongue too. No wonder, they are tempted to let their hands help, because their tongue proved to be insufficient. These quivering fingers are very effective, they overrule the sense of every spoken word.

But they make - sorry to say - a gesture of obscenity. Of course, the obscenity is somehow childish, or boyish, but it throws a rather doubious light on these speakers. The gesture is understandable, and I grant anyone the right to use them. But let me make this very clear: I would never use them myself. This is a personal confession. And there won't be any reason to come back to this point.

Nevertheless, the personal statement is useful. Today's gestical way of handling quotation marks is just a private matter. The historical way to cope with written quotational material was no private matter at all. On the contrary, once upon a time quotation marks were used in order to officially point out words of literary and divine authority.

1.1 Quotation marks enclosing their items

In our century we are accustomed to seeing quotation marks at the beginning and at the end of a certain item of text. Together the initial and the final sign form a pair of signs. The enclosed item may be short or long, the number of marks consists of two signs only.

Their position is in accordance to the incidental placing in a line system: most often someplace in the middle of the line. Though sometimes initial marks can be placed at the beginning of lines and final marks at the end of lines, their frequent line position is the middle.

Their function is to change the over-all-meaning of the enclosed item.

1.1.1 Standard marks in editing

Quotation marks with their characteristic twin appearance and placement in the letter area were unknown before 1650. Yet during the last century they were declared standard in orthography and typography in many Western countries. Unfortunately this standard became a norm for text editing too, even older texts. Therefore anachronistic norms quickly filled new editions of old literature. Old texts were turned upside down:

Quotation marks were added where there had not been any, and - what is worse - the marks of the original disappeared systematically from the texts.

Because of the eager activities of last century's editors many of us have never seen the old type of quotation mark. In order to have a look at them you have to open old books yourself. Facsimiles and photos can be helpful too.

Plate 1: Evangelia S. Burchardi, 690-710, Würzburg, Univ.-Bibl. M.P.Th.F. 68, page 96v. - Luc. 3, 4-7. - A marginal chain of s-shaped signs escorts 9 lines containing a quotation from Esaias 40, 3-5.


2. The quota sign escorting its item

It is time to introduce a name for the (so called) quotation marks before 1650. The total difference in placement - possibly many differences in function and shape too - makes it desirable to call them by a special name. Therefore from here on we make a terminological distinction between quotation mark and quota sign.

By quota sign we mean the mark before 1650.

We continue to call the signs after 1650 quotation marks .

Now it is easy to focus on the purpose of this paper: to describe the difference between quotation mark and quota signs. The emphasis will be the placement of the signs rather than their shape and their function.

2.1 The placement in the margin

The most typical feature of the quota sign is its position on a page or a double page. You do not find it inside of the letter area (the area normally being written or printed on). It is always placed in the empty space to the right or to the left of the letter area, i.e. in the margin. As a marginal annotation it has the advantage of high visibility. The quota sign is an eye-catcher.

It shares position with and sometimes struggles against other marginal items like side titels, marginal glosses, reader's remarks and - important for the Middle Ages - the Eusebian reference codes.

Through its visibility it opens up certain items or passages of the text for special reading. This is true even for a prospective reader just browsing through a codex: turning pages he is suddenly stopped by a quota sign, he reads the marked passage, perhaps becomes enthusiastic, and, finally, feels motivated to read the whole book.

2.1.1 Unchanged in one and a half millenia

The collateral position enabled the quota sign to remain unchanged for many centuries. Inspite of major changes in the system of writing (uncial, minuscul, textura etc.) the quota sign - being an outsider - was untouched.

And it is, of course, not by chance that the quota sign finally had little chance to survive, when it - around 1650 - lost its collateral position and was printed on the border of the letter area. Thereby it sacrificed much of its visibility and could no longer function as previously. The shortcoming was gradually compensated for by distributing quotation marks in the letter area. It did not take much time before the quota sign became useless. And apparently it is only due to its long tradition that the chain on the border was not expelled at once. One sees it very often in texts from 1650 to 1800 and scarcely later on. Leibniz used it, and so did Voltaire and Lessing. Goethe, who despised it in his youth, used it in his old days, just when others had abandoned it.

2.2 Relation to lines

The quota sign is collateral to the letter area. But each of its parts is collateral too, because they are aligned to the text lines. The quota sign therefore displays a double collateral relation: to the letter area on the whole, and to each of its lines.

Furthermore, all parts of the quota sign hold the same distance to the letter area and are placed directly below each other. This way they build up the whole quota sign to a vertical unit of high regularity, a unit pleasant to look at. By these features - the collateral placement and the collineal subdivision - the quota sign is able to guide the reader's eye. Its function is a very special one:

The (collineal) quota sign introduces the reader to some special lines of the text.

Properly speaking, it doesn't direct one to the text itself or to the words of the text, but rather to the lines bearing the words and text. To be a line pointer, this is its only function. You may call it a line index.

An author or a scribe marks a text because he wants to guide the reader to special lines. Lines of particular interest for the author, the translator, the interpreter can be emphasized this way.

This sounds simple. But the marking of lines becomes highly complicated if you study examples and ask yourself what may have motivated their marking. The full-fledged field of motivations for line marking can only be indicated here.

Lines may contain some spurious words, which the author warns the reader for. A translator is in doubt about a certain passage in his own translation, so he wants to notify the next translator. In both cases the quota mark is an editorial signal. Other lines may contain something heretical, so the author protects his reader with the warning marks. Other lines may contain the sacred word of the Lord, which the author decorates. Here the quota sign becomes something like a gloria. It was Cassiodorus in the 6th century who mentioned a sign not unlike a quota sign (Latin double-comma-type). He tells us that the sign defends certain passages of the sacred law:

We have also marked idioms of the sacred law, that is, expressions peculiar to it, with the character of this sort p^p, in order that these words, wherever found, may be profaned by no presumption. [Divine Letters, book I, XXVI,2]

2.3 A metaphor

The special display of the quota sign makes it irresistable to compare it with similar structures. The quota sign is, for example, like a pendulum, like a marshal's beaton, like a string with regular knots, or like a flock of geese marching in Indian file (the Germans talk about Gänsefüßchen ). Another comparison reds:

The quota sign is like a hanging chain and its links.

This is meant figuratively, as figurative as the catena -model in medieval educational theory (Carruthers 1990: 5-6). From now on we may use the word chain instead of quota sign .

Wherever one finds chains, one sees they function as signs escorting a passage of text, and they do it in the margin. How this escort transposes the marked text has to be uncovered by semantic analysis. Through the marginal escort it has all requirements to be in command of the marked passage.

3. Ptolemean origin (hypothesis)

Some centuries before the first appearence of quota signs a normative system of marginal signs had been developed in the Greek colony of Egypt. The Musaion in Alexandria, initiated by Ptolemaios II., was the very place where scolars from all over the Greek empire gathered in order to collect and collate manuscripts (Reynolds,Wilson 1974: 5-15).

3.1 Corrections signs

One assumes that Zenodot from Ephesus (about 270 B.C.) developed the first normative correction signs, and that Aristophanes from Byzanz (about 200 B.C.) took over and enlarged their number. We know that Aristarchos from Samothrake (about 160 B.C.) systemized six of them forming a correction system comparable to our set of correction signs for proof-reading (Pfeiffer 1968: 112-115) .The number of signs steadily increased. When Isidor from Sevilla counts them some eight centuries later he specifies as many as twenty-six known as sententiarum notae (Etymologia, I. 21.). Some of these signs we still apply today and use their old names, such as the paragraph and the asterics .

3.2 The sign called diplê

There is no doubt that already Zenodotos introduced a sign which four hundred years later became an element of the quota sign, its form being / > / , its name diplê . Like the other correction signs in the Homeric text critism it was written in the margin and on level with a certain line. Its function was rather general, calling attention to something special in the line, a splendid coinage of the author, an orthographic mistake of a scribe, a missing metrical foot, or other rather small items, but no quotations, as far as we know.

There is more evidence for the Ptolemean origin theory of the quota sign. All the early examples of chains we find on Greek papyrus were dug up in Egyptian soil. Pagan, Jewish and Christian documents marked by quota signs have Egyptian provenance. They may be copies from Platon, (1. on the list), from the Jewish philosopher Philon (2. on the list), or the Christian dogmatist Irenaeus (3. on the list).

There are, however, missing links of manuscripts handed down from the 3nd century B.C. to the 2nd. century A.D.. The theory of the origin of quota signs in the Alexandrian text criticism remains hypothetical.

4. A list to prove continuity

The extravagancy of writing and printing in the margin was disliked by many scribes and printers from the Ptolemean times throughout the Middle Ages up to the period of the Enlightenment. It is not surprising then that most texts from 200 to 1650 have no quota signs at all.

This has a certain advantage if you want to mark off the whole field of chain application. If there is only a relatively small amount of texts marked by quota signs, it is a realistic project to list them. Such a listing need not even be complete from the very beginning. A rough diachronic sketch might be helpful in a preliminary survey. This will be attempted here. Later on the rich synchronic parallels will have to be filled in.

Several scolars have started such a list, but limitations of either time, language, or provenance have prevented a general survey. Two of them have to be mentioned here.

For three decades E.A. Lowe described antique and medieval codices always mentioning shape and placement of quota signs as a typical feature of manuscripts (Lowe 1934-1966); yet his work is limited to Latin texts and furthermore to texts up to the year 800 A.D. only. This is also true of his later list (Lowe 1972).

In 1961 Patrick McGurk presented very early Greek and Latin texts with quota signs in an article in Scriptorium (Tome XV: 3-13), but he limited himself to manuscripts found in British libraries.

4.1 Diachronical documentation

The following list originates from the question of whether or not it is reasonable to talk about a tradition of quota marking. Is there at least one chain marked text in each of the first 15 centuries? This would be a minimum requirement for stating a existing tradition.

The list below doesn't cover more than six languages: Greek and Latin as basics for general communication (from the 2nd c.) and some national languages like Gothic (5th and 6th c.), German (from the 8th c.), French (16th c.), and Italien (17th c.). German texts have been followed more closely from the 8th to the 18th century. There are two excellent paragraphs in a doctorial thesis from Paris written by Nina Catach (Catach (1968: 79- 81, 299-301) doing this for French usage of the 16th century.

Nearly all items on the list are taken from second hand paleographical literature and its increasing amount of photographic evidence. Only the collections of the last 100 years need to be named here (in chronological order):

Chatelain, Émile 1884: Paléographie des classiques latins, Paris

Könnecke, Gustav 1894: Bilderatlas zur Geschichte der dt. Nationalliteratur, Marburg (2nd ed.)

Enneccerus, Magda 1897: Die ältesten deutschen Sprach-Denkmäler, Frankfurt /M.

Facsimiles of Ancient Manuscripts 1903-1924, 6 vol., (New Paleographical Society ed.) London

Karabacek, Joseph, Ritter von 1910: Monumenta palaeographica vindobonensia , Leipzig

Rand, Edward Kennard 1929: A Survey of the Manuscripts of Tours, 2 vol., Cambridge/Mass.

Jones, Leslie Webber 1932: The Script of Cologne, Cambridge/Mass.

CLA = Codices latini antiquiores , 11 vol., 1934-1966 (Elias Avery Lowe ed.), Oxford

Bruckner, Albert, 1935-1978: Scriptoria medii aevi helvetica, 14 vol. , Genève

Seider, Richard, 1967-1990: Paläographie der griechischen Papyri, 3 vol., Stuttgart

Seider, Richard 1972-1981: Paläographie der lateinischen Papyri, 3 vol. Stuttgart

Metzger, Bruce M. 1981: Manuscripts of the Greek Bible, New York

Turner, Erik G. 1987: Greek Manuscripts of the Ancient World, London

There are at least as many monographs (like Berliner Klassikertexte or Der Wulfila der Bibliotheca Augusta ) from which material is compiled.

I had only a few chances to work with original manuscripts. Thanks to the generosity of the librarians in Engelberg, Freiburg/Br., Fulda, Gothenburg, St. Gallen, and Wolfenbüttel I was allowed to spend days or weeks pouring over parchment codices. In doing so I found that the paleographical second hand literatur has quite a few missrepresentations and missinterpretations in respect to quota signs.

In the list there are eight columns presenting details on the placement of quota signs. The columns A and A* represent the left and right margin of a page, B and B* stand for the left and right border of the letter area, C and C* mean any place in the letter area.

4.2 List on occurence and placement





















































































































































Isidore Sev.












Isidore Sev.




















































































by hand























F Tory

















































































































































































time: the century a manuscript was rewritten,

or the year it was printed / reprinted.

The time for reproduction seldom coincides with the time of the original's production.

author/title: Paleographical names are used for texts handed down anonymously.

languages: F = French, G = Greek, Ge = German, Go = Gothic, I = Italian L = Latin

place: location of the library holding the manuscript today (for handwritten books)

or location of the publishing company (for printed books).

4.3 Conclusions from the list

The list contains handwritten texts from the 2nd to 16th century (1.-20. on the list) and printed texts from the 16th to 20th century (21.-36.).

Based on the list one can say:

Quotative marks have some type of tradition 1800 years.

Marks in the letter area were exceptional before 1650.

The tradition of real quotation marks started as late as 1650.

4.3.1 A certain tradition

From the 2nd to the 20th century some type of quotation marks or quota signs were used. The tradition of using quota signs was fully established in the late classical writing of the Greeks. Evidence may be found in pagan (1.) and religious (2.) papyrus and mainly in parchment copies of the New Testament (3.-5.,7.,9.-11.) and its commentaries (6.,8.,13.-18.). The apogee of quota sign tradition can be found in the incountable copies of Johannes Chrysostomos' Homilies , for nearly all of them bear quota signs in diplê-shape throughout the Middle Ages , even in their Latin translations (not on the list).

In Latin writing comparable long-term habits are difficult to find. Of course, there were some local writing customs, e.g. in some Gothic scriptorium of the 6th century (5., 7., 8.), or at Tours and Cologne in the Carolingian period (not on the list), but it would be exaggerated to talk about a graphical norm.

There are periods where the quota sign apparently became outmoded, e.g. in the beginning of the national literatures (13. being an exception) and in the early period of printing. Nobody knows what would have happened , if Aldus Manutius had not reintroduced quota signs in his editions from 1502 to 1514 . He annotated classical texts (21.) marking sentences, proverbs, and quotations too. Because he sold copies all over Europe and because other printers imitated his chain marking - first in Florence 1522, then in Basel 1523, and finally 1529 in Paris by Geofroy de Tory in his bestsellerChamp Fleury (22.) - the old chain was rescued and its reintroduction was not to be stopped.

4.3.2 Experiments to mark in the (middle of) lines

The margin is the place for the marks up to the 17th century. But experiments to mark texts even inside of the letter area were sometimes going on (8.,14.), though never prevailing.

The early printers (21.- 24.) imitated the scribes and held on to the marginal notation for two more centuries. The big change in placement happened as late as around 1650 and not around 1450 as one could expect.

4.3.3 The preference of the left margin

The quota signs appear mostly in the left hand margin. The right hand margin could never really compete. Obviously the left margin is preferred, because a single page concept decided all placement questions. Of course, in a double page concept the right hand margin on the right hand page becomes the symmetrical counter part of the left hand margins on the left hand page. They become outer margins and consequently both could bear the quota signs. This is true for written texts (6. ?, 17., 20.) as well as for printed texts (23., 27., 28., 32., 33.). Geofroy de Tory (22.) even chose the inner margin for quota signs reserving the outer ones for side titles.

5. Other features of enclosing and escorting marks

Finally we should look at some changes in shape and function taking place simultanously with the changes in placement.

5.1 Variation in shape

As to the shape of every link in the chain the variety is tremendous. There are, however, three forms with a long tradition: the angular form / > / or / >> / called diplê . It is used in Greek and Gothic manuscripts from Ptolemean times thoughout the Middle Ages. At first there was one angular sign per link of the chain(1. - 6.), later on there were often two of them side by side.

The wavy shape /? / or /?? / in Latin, Gothic and German manuscripts used mostly in religious writing from the 6th century (8., 9.). throughout the medieval times (10.-14., 16., 17., 19.), but never in print.

The comma shape / , / or / ,, / being applied sometimes round 800 (13.). Later on it became the standard form in early printing (21.-34.). The reverse shape / / or / "/ was used additionally in the right hand margin (22., 23., 27.) or on the right hand border (28., 29.), and finally became the opening signal for the usual quotation mark in British typography (called reversed / inverted comma ).

In today's printing there are many different standard shapes in different countries. Even in modern times there are important changes in shape , for example the transition from /"..."/ to /«» / in Swiss typography and to /»« / in German typography during the last 15 years.

5.2 Variations in function

There are many different functions of the quotation mark and even more for the quota sign. The general missinterpretation of the quota sign as a quotation mark cannot be analysed here, but some special points may be mentioned.

1. Looking into the early copies of the Gospels (in the Codd. Sinaîticus, Vaticanus, Alexandrinus) one soon finds that quotations from the Old Testament are sometimes marked, sometimes not. Consequently these chains were not meant to mark quotation as such.

2. When Aldus Manutius in 1502 introduced quota marks into printing he marked quotable sayings of Sophocles, but not any quotations (21.)

3. Furthermore, novellists started as late as 1680 to mark verbatim rendering of their characters in novels (28.), as did historians when referring to verbal documents (Pufendorf 1696, 1697).

There are many more functional possibilities for quota signs (Kluge 1993). Here is one more aspect. Thinking of the ceremonial context of recited texts quota signs might have been the cue for passing the word to another speaker or speakers (the choir in church or on stage), or at least for a change of pitch by one and the same speaker. This way quota signs could have been pragmatic signals for voice distribution or means of voice registering (Klockow, 1980: 7-17).

The whole field of quotive differention by semiotic means is not yet really marked off. Research on this field needing interdisciplinary teamwork has just started.


Catach, N. 1968: L'orthographie française à l'époque de la Renaissance, Genève.

Carruthers, M. C. 1990: The Book of Memory, Cambridge.

McGurk, P. 1961: Citation Marks in Early Latin Manuscripts

in: Scriptorium, Tome XV: 3-13 (plus four plates), Anvers, Amsterdam.

Klockow, R. 1980: Linguistik der Gänsefüßchen, Frankfurt/M.

Kluge, H. 1993: Über die Funktionen der Anführungszeichen im 17. Jahrhundert, in: Sprachgermanistik in Skandinavien, pp. 81-92, Göteborg.

Lowe, E.A. 1934-1971: Codices Latini Antiquiores, A Paleographical Guide to Latin Manuscripts Prior to the Ninth Century, Oxford.

Lowe, E.A. 1972: Paleographical Papers 1907 - 1965, Oxford.

Pfeiffer, R. 1968: History of Classic Scholarship, Oxford.

Reynolds, L.D. and Wilson, N.G. 1974: Scribes and Scholars, Oxford.

Weyers, Chr. 1991: Zur Entwicklung der Anführungszeichen in gedruckten Texten, in: Zeitschrift für Semiotik, vol. 14, 1-2 : 17-28.




Marlene Landsch

Free University of Berlin






Before we can begin a discussion on the meaning of visual signs, one point has to be clear: we are not dealing with a homogenous concept. This is demonstrated in the communication process.

The interpretation of visual signs involves not only a deciphering of meaning on the basis of specific codes and styles, but also an interpretation of meaning on the basis of the specific world knowledge of the recipient. This observation rests on the view that the meaning of a visual sign derives from two complementary sources:

- On the one hand, it is assumed that the viewer who is familiar with the code finds a meaning (signifikant) in the sign which is a property of the sign itself.

- On the other hand, it is assumed that "pictures can't say ain't." (Worth)

Meaning is ascribed to visual signs. This type of meaning assignment is dependent upon socio-cultural, psychological, and situation-specific contexts in which the viewer is embedded upon stereotypes and conventions which he has internalized.

The information which a picture imparts is, thus, not only static i.e. dependent upon the given sign, but also dynamic, i.e. dependent on that which we actively supplement as interpretation. The viewer of a picture integrates the information from the visual imput with his world knowledge and forms a unified, resolved manner of looking at the "situation".

When they are combined, pictures, as well as language expressions, may obtain a total meaning which is more than the sum of the meanings of their parts. An additional meaning can be assigned indirectly to a picture in a series which it does not possess in isolation.

The purpose of the present investigation is to demonstrate factors of interpretation relating to the special organization of individual pictures or of a sequence of two. We will be less concerned with the style of the Peanuts paradigm chosen than with the visual code common to pictures of different styles. That is, I am less interested in questions of whether the picture is an etching, watercolor, or ink drawing, or whether it is "realistic" or "unrealistic". What I will discuss is an aspect common to all pictures: the spatial code.

The spatial code requires a characterization of three-dimensional objects in two-dimensional space. Such a portrayal allows for varying perspectives:

certain objects may be emphasized, others deemphasized, and individual parts may be recombined under contextual influence.

Spatial presentation in the material I investigated is static, i.e. even persons who are perceived by the viewer to be moving are depicted in an unaltered position. What is important is the decisive role of perspective in the interpretation of movement, which may be seen as goal- or source-oriented. This observation is valid in particular for nonisomophic perspectives, and in this study we will examine whether it is also true of isomorphic perspectives, i.e. for those which can be reversed through the geometric operations of rotation and reflection. How is interpretation influenced by the following factors:

1.whether a person who is shown in isolation is moving from left to right or right to left?

2.whether a person who is shown within a picture frame (i.e. border) is moving from left to right or right to left?

3. whether one picture of a sequence shows movement and the other does not? The individual sequences also vary with regard to a systematic change in the left/right relations; the order in which the pictures are


In order to answer these questions, an empirical study was conducted which consisted of three phases separated by eight-day intervals. In Phase 1, subjects were presented two isolated portrayals of movement in which left-right-relations were varied; in Phase 2, the same pictures had been placed within borders; Phase 3 involved the interpretation of eight picture pairs which varied systematically both in their left-right-relations and order of presentation.

Results were obtainded from a population of 148 randomly-selected native German speakers of which 70 were men and 78 women between the age of 18 and 64. 48 subjects were students. 36 subjects were left-handed as defined with regard to left dominance of the motor apparatus, and 28 also wrote with their left hand.

In addition, 41 Arabic speakers also participated. 25 received their instructions in French. However, 16 of 41 spoke Arabic only, and so their descriptions were taped and later translated into French.

The relatively high percentage of left-handed subjects was intentional, given the hypothesis that left dominance in the motor apparatus might correspond to a left dominance in picture interpretation (and vice versa). Arabic subjects were chosen because it was also hypothesized that the interpretation of actions moving from left to right or right to left might be dependent upon internalized cultural practices such as reading and writing: from left to right in our culture, but from right to left in the Arabic world.

Investigations in the Arabic area will be repeated in the near future: the most interesting results were obtained from the 16 illiterates. Unfortunately, the sample size was too small for a significant determination of probability. Three subjects responded in a particularly noteworthy manner: rather than reconstructing the picture sequences as sequences, they operated with statements such as That guy there is going, or That guy there is sitting.

8 subjects mixed the contents of the pictures with personal experience, providing "personal accounts" which have yet to be evaluated more thoroughly, but which will certainly be very informative regarding to the function of personal background in interpretation. For the above reasons, data from the Arabic area are only used supportively.

Picture 1

(Adapted from: Schulz, C. (1979) Peanuts. The Misfortunes of Charlie Brown, 19. Hodder and Stoughton, London, Sydney, Auckland, Toronto.)

In Phase 1, subjects were shown two pictures. The content consisted of a moving person (Charlie Brown). No further visual context was present, and the second picture was a mirror image of the first. Because the perspective requires that the viewer establishes a relation between himself and the object he is looking at, the pictures were placed together on one piece of paper 8 1/2" by 11 1/2" in order to prevent errors. The drawings were located in the center of the sheet, one above the other. Instructions were repeated for each picture and were phrased as follows: Which verb do you associate with this picture: kommen (Eng. 'come'; Fr. 'venir' or gehen ('go'; 'aller')?

The pictures did not contain what Piaget (1975:270) calls a "dominating characteristic" which might provide the viewer with a point of orientation. As a result, only two expressions in the system of topological relations could be used to designate the direction of movement: left and right.

In response to the picture depicting movement from left to right, 130 (of 148) subjects (87.83%) decided on the expression kommen. 6 asked if they could also say Der geht irgendwohin, 'He's going somewhere'. 12 chose the expression gehen. 6 subjects were unable to make a decision.

A linguistic analysis of kommen und gehen shows that both verbs reflect what we can call terminated movements:

1. Kommen is goal-oriented, designating movement towards the speaker, and gehen source oriented, designating movement away from the speaker (Vernay 1974:136 ff.).

2. The speaker may transfer the orientation point from his own location to one with which he identifies (Fillmore 1971), e.g. A is speaking on the telephone with B about C, D, and E. A says: "C is going from D to E." In this case, A identifies with D. Or A says: "C is coming from D to E." In this case, A identifies with E.

3. The expressions hingehen 'go' (to a certain place; somewhere) and herkommen 'come' (from a certain place) are only more complex forms of gehen and kommen (Vernay 1974:137).

A total of 22 persons were indecisive about whether they prefer kommen or hingehen and gehen or herkommen; their statements are therefore not included in the further evaluations of Phase 1. Subtracting the judgments of the six subjects who preferred hingehen from the 130 who decided on the expression kommen for picture 1, we find that 83,78% chose kommen for the picture showing movement from left to right.

Deducting the responses of those 16 subjects who preferred the expression herkommen from the 125 who selected gehen for picture 2, we find that 73.64% chose gehen for the picture showing movement from right to left.

One might be tempted to try to explain the above results by seeking a point of orientation which would influence the viewer in converting left-right movement into kommen and right-left movement into gehen. However, as I indicated, no such points are present in the pictures used, and we must therefore look for an explanation elsewhere. I suspect that the intellectual egocentrism present in the meanings of the verbs kommen (...zu mir hin, 'toward me') and gehen (...von mir weg, 'away from me') also play a decisive role in the perception of the picture.

In Phase 2, subjects were presented with two pictures of individuals moving within the reference of a frame, or border. The second picture was a mirror image of the first. The pictures were arranged vertically, and it was immediately recognizable to the viewer that with regard to their respective frames, left-right movement was drawn in the left portion of the picture, and right-left movement in the right portion.

The instructions given to the subjects (for each picture separately) were:

In your opinion, which of the following sentences corresponds best to

the picture:

-- Der Mann geht, 'The man is going';

-- Der Mann kommt, 'The man is coming';

-- Der Mann geht irgendowhin, 'The man is going somewhere';

-- Der Mann kommt irgendwoher, 'The man is coming (from) somewhere';

-- Der Mann geht irgendwoher, 'The man is going (from) somewhere':

-- Der Mann kommt irgendwohin, 'The man is coming somewhere'?


Picture 2.1 Picture 2.2

(Pictures 2.1 and 2.2 were adapted from: Schulz, C. (1979) Peanuts. The Misfortunes of Charlie Brown, 19, Hodder and Stoughton, London Sydney, Auckland, Toronto.)

In Phase 1, pictures contained no "dominating characteristic" which would provide the viewer with a point of orientation. They were interpreted from left and right in the system of topological relations, the main factor being the perspective of the viewer, the relation he establishes between himself and the picture. Pictures of Phase 2, however, do exhibit a characteristic which can influence orientation: the frame. In this case, the viewer not only has his individual perspective, he is also given a spatial point of reference, the division of space provided by the picture. The viewer no longer relates to the figure alone, to the figure in the frame.

Phase 2 results: When viewing picture 1, showing Charlie Brown moving from left to right, subjects made the following judgments:

1. 2 (of 148) (1.35 %) for Der Mann geht;

2. 8 (of 148) (5.4 %) for Der Mann kommt;

3. 102 (of 148) (68.91 %) for Der Mann geht irgendwohin;

4. 28 (of 148) (18.91 %) for Der Mann kommt irgendwoher;

5. No one for Der Mann geht irgendwoher.

6. 8 (of 148) (5.4%) for Der Mann kommt irgendwohin.

When viewing picture 2, showing Charlie Brown moving from right to left, the following decisions were made:

1. 12 (of 148) (8.1 %) for Der Mann geht;

2. 6 (of 148) (4.05 %) for Der Mann kommt;

3. 21 (of 148) (14.18 %) for Der Mann geht irgendwohin;

4. 108 (of 148) (72.97%) for Der Mann kommt irgendwoher;

5. No one for Der Mann geht irgendwoher;

6. 1 (of 148) (0.67 %) for Der Mann kommt irgendwohin.

A linguistic interpretation of the expressions kommen and gehen was made above. In order to interpret the expressions hingehen, hergehen, hinkommen and herkommen, I have utilized the vector system of Vernay (1974:137). In this analysis, the particle hin indicates a direction of movement which begins at that point where the communicator (in this case viewer of the picture) is located „ici du moi" (Lo) and ends at a point where the communicator is no longer located „non-ici du moi" (-Lo) (Vernay 1974:56). The movement indicated by the particle her starts from a point at which the communicator is not located, "non-ici- du moi" (-Lo) and ends at a point where he is "ici du moi" (Lo)(Vernay 1974:56). The direction of movement of the particle hin thus goes from (Lo) to (-Lo), that of the particle her (-Lo) to (Lo).

When the pictures were presented, all of the subjects remembered "having already seen something like that before." However, after the experiment, none of them could say, without being helped, what the difference was between Phase 2 pictures and those shown eight days earlier. Individual score sheets were coded for each subject, and so a comparison can be made between the responses of the first and second phases.

Interpretation of individual results: Picture 1, Sentence (1), Der Mann geht

The two subjects who chose this sentence originally interpreted movement from left to right as kommen. In Phase 2, however, their orientation point is the spatial reference supplied by the picture. The person depicted is going away from the left border, and the rightward point of orientation originally required for a kommen-interpretation has been shifted to the left.

Picture 2, Sentence (1), Der Mann geht.

In Phase 1, all 12 subjects interpreted the movement from right to left as gehen. However, although the viewer's own perspective is supported in Phase 2 by the spatial distribution of the picture - the person is moving away from the right border - all 12 subjects interpreted the picture from the left side.

Picture 1, Sentence (2), Der Mann kommt. In Phase 1, all 8 subjects interpreted the direction of movement from left to right as kommen. In Phase 2, the perspective of the viewer is supported by the spatial division of the picture - the person is moving away from the left border - yet the viewer imagines himself on the right.

While all 6 subjects interpreted right-left movement as gehen in the first phase, the point of orientation has now moved to the left border, which the person depicted is coming towards. The point of orientation of the viewer is thus on the left side.

Picture 1, Sentence (3) Der Mann geht irgendwohin. In Phase 1, 86 of the 102 subjects interpreted movement from left to right as kommen, 6 asked if they could say Der geht irgendwohin, 'He's going somewhere', and 10 decided on the expression gehen. Point of orientation for interpretation in the second stage is both the spatial division of the picture (i. e. the left border, which the figure is "going " away from) and the perspective, i.e. the relation which the viewer sets up between himself and the picture: the viewer imagines himself in a position on the left edge of the picture according to the analysis of the particle hin as showing movement from (Lo) to (-Lo). Whereas point of orientation for kommen would be situated on the right, for Der Mann geht irgendwohin it is shifted to the left.

Picture 2, Sentence (3) Der Mann geht irgendwohin. In the first phase, 17 of 21 subjects interpreted right-left movement as gehen, and 4 as kommen. Reference in the second stage is provided by both the spatial division of the picture, the right edge from which the figure is "going" away, as well as by the perspective established by the viewer: the viewer imagines himself on the right edge of the picture, which conforms with the analysis of the particle hin as an indicator of movement from (Lo) to (-Lo).

Picture 1, Sentence (4), Der Mann kommt irgendwoher. In Phase 1, 26 of 28 subjects described left-right movement as kommen, and 2 as gehen, an interpretation maintained in Phase 2. The viewer imagines himself in a position at the right edge of the picture, which corresponds to the analysis of her as indicating movement from (-Lo) to (Lo).

Picture 2, Sentence (4), Der Mann kommt irgendwoher. Of these 108 subjects, 89 chose gehen for right-left movement in Phase 1; 16 asked if they could say Der Mann kommt irgendwoher; 3 chose the expression kommen. The point of orientation thus changes: originally on the right, as indicated by the selection of gehen, reference shifts and the speaker imagines himself in a position on the left border. This parallels the analysis of her as portraying movement from (-Lo) to (Lo).

Picture 1, Picture 2, Sentence (5), Der Mann geht irgendwoher. All German-speaking subjects were native speakers, and they accurately perceived this sentence as "impossible". It is indeed ungrammatical: the movement reflected in the verb gehen is form (Lo) to (-Lo), whereas in the particle her it is form (-Lo) to (Lo), and the two directions are mutually exclusive.

Picture 1, Sentence (6), Der Mann kommt irgendwohin. Although the direction of movement of the verb kommen from (-Lo) to (Lo) and that of the particle hin from (Lo) to (-Lo) exclude one another, all 8 subjects interpreted picture 1 with sentence (6). That this combination of kommen and hin was not perceived as ungrammatical is apparently due to the fact that the goal-orientation of the verb kommen is less strong than the source-orientation of the verb go. All 8 persons had decided on the verb kommen in the first phase. Interestingly, the viewer's point of orientation is now both the right border of the picture, as we see from the choice of kommen, as well as the left, as we see from the choice of hin.

Picture 2 Sentence (6), Der Mann kommt irgendwohin. These subjects decided on the verb gehen in Phase 1. With regard to their position as viewers, the same factors apply in this case to the 8 above mentioned subjects who interpreted picture 1 with sentence (6).

When discussing the results of Phase 1, I indicated my suspicion that intellectual egocentrism plays a decisive role in the association of the verbs come and go with movement in a given direction. Moreover, the position which the viewer imagines himself when interpreting Phase 1 pictures is definitely on the right. However, when spatial coding is added to the information of a picture, as in Phase 2, the viewer's point of orientation makes a significant shift from the right to the left border. Spatial coding thus precludes the free choice of an orientation point.

In Phase 3, subjects were shown a set of pictures. Materials consisted of 8 picture sequences with 2 pictures each. The two were the same in each sequence, but the sequences differed in the ordering of the two pictures and in the systematic change in their right-left relations. With regard to orientational features, the pictures of each phase of the experiment can thus be characterized as follows: those of the first, displayed no dominating characteristic which might serve as a point of orientation; those of the second, contained spatial coding; those of the third were also ordered. Directions for Phase 3 were

1. In the following exercise, you will be given 8 pairs of pictures on

separate pages. The pictures in each pair are related in content (to each other, but not to other pairs).

2. Determine the relationship between each pair of pictures and describe it in two or three sentences on the same page.

Picture 3

(In the original German instructions, the verb darstellen, 'represent', was used for determine. The redundancy between darstellen and beschreiben, 'describe', proved necessary because some of the subjects thought that they were to assign speech bubbles to the respective pictures, i.e. that they were to construct direct dialogue).


A textlinguistic methodology was used to analyze the sentences of individual sequences, and the results were summarized with regard to three questions:

1. What effect does figure movement have on statements about pictures 3 (left-right) and 4 (right-left)?

2. What is the influence of left-right rotation in picture 1 and of right-left rotation in picture 2?

3. How does ordering within a sequence influence contextual interpretation? Given that individual and texlinguistic results correspond in 78 % of all cases (based on a sample size of 148), individual figures for the former will be omitted in the following discussion. The 22 % error rate arose for the most part because 1) certain subjects were not able to perceive anything beyond the Peanuts paradigm itself, and 2) other subjects were unwilling, and apparently also unable, to explain the "same" pictures eight times and to put what they had seen into writing.

With regard to our first question, whether it makes a relevant difference in interpretation if a figure in the context of a seqence is moving from left to right or right to left, I found that in 1.3 and 2.3 the interpretation of picture 3 is source-oriented and can be paraphrased by the word weg von, 'away from'. In 3.1 and 3.2, the interpretation of picture 3 is goal-, as well as source-oriented, and can be paraphrased by the words hin zu (roughly 'toward'). In 1.4 and 2.4, the interpretation of picture 4 is goal-oriented and can be paraphrased by the words hin zu. Picture sequences 4.1 and 4.2 are rather unique and will be discussed in more detail below.

The motional particles hin zu and weg von also correspond to the verbs kommen and gehen and to their paraphrases (Fillmore 1971:369 ff.; Vernay 1974). Thus, the visual coding of a movement process from left to right can be correlated with the particle hin zu, and the visual coding of a process from right to left can be correlated with the verbal coding of weg von.

The relative instability of the picture interpretations seems surprising when one regards pictures 3 and 4 more closely. We see the front of a house, a shrub, a step, and a movement process in which a person (Charlie Brown) is going down the step (in a direction) with his back to the house. Moreover, none of these features - except for the step - extends beyond the middle of the picture. Before our experiments, I therefore suspected that this clear spatial coding would provide a focal point of interpretation, even within a further context: the person isleaving the house, thus weg von, 'away from'. Because the number of the above elements is constant, whether the picture is rotated from right to left or left to right, it seemed logical to assume that subject verbalization of the relations in each pair of pictures would only yield von weg.

To help explain the viewer's interpretations, I will quote passages which, when reduced to their linguistic content, are representative of the 78 % of the answers which coincide.

The following explanation was given for sequences 1.3 and 2.3: "The boy has apparently got a date (or some kind of appointment). He's using the time he has left to read, then he looks at the clock and sees that it's time to go. In the second picture he's going out of the house."

The inside space of picture 1 and the outside front of picture 3 serve as brackets which connect the two. With this spatial coding the persons are perceived as the same indiviual, and the orientation point is in each case picture 1.

For sequences 3.1 and 3.2, there are essentially two possible contextualizations, depending upon whether the person depicted is perceived as one or as two individuals. If he is seen as one person, then the direction of movement in the first picture, 3, is described with kommen: "The boy is coming home. He sits down and wants to read a book." The Weggehen aus dem Hause, 'going out of the house', as it is shown in the first picture of the sequence, is ignored.

However, if the figures are perceived as two different people, then the movement in the first picture, 3, is described with gehen: "One of the boys is leaving the house. He wants to visit the other one, who's waiting for him."

In sequences 3.1 and 3.2, the interpretation of the relationship between pictures 3 and 1 is thus determined by picture 1.

Verbalizations of sequences 1.4 and 2.4 went as follows."The boy is waiting for a visitor and looking expectantly at the clock. He doesn't know that the visitor is already on his way and will arrive soon." The figures in 1.4 and 2.4 are thus perceived as 2 different people, and the context is explained with picture 1 as focal point. Since this picture also appears first in sequences 1.4 and 2.4, the same order is maintained in the interpretation.

We can thus conclude that it makes no relevant difference in picture sequences whether a person is moving from right to left or left to right. In other words, interpretation of the direction of movement is in each case dependent upon context.

In question 2, we asked whether left-right rotation, as in picture 1, or right-left rotation, as in 2, would have an effect on subject statements. In this case, I found that rotation had no influence on the interpretation of cognitive content, although it had on the interpretation of affective content. When affective content was verbalized, it was almost always judged more positively in picture 1, with the person sitting from left to right, than in picture 2, with the person sitting from right to left. In the first case, the German expression gemütlich (here roughly translated as 'cozily' or 'comfortably') was often used, but in the second case ärgerlich, 'angry' or 'annoyed':

"The boy is sitting cozily in the chair. It's only five to six."

for picture 1, versus "The boy is angry. It's already five to six."

for picture 2.

A number of structural elements found in the pictures could also be discussed and related to the above findings: e.g. the ordering of the objects in the room; the time on the clock; the position of the lamp. However, only an empirical study which varied each of these elements systematically could elucidate the affective potential of the pictures.

With regard to question 3, how the ordering of the pictures influences interpretation of context, I found that pictures 1 and 2 always control interpretation, regardless of their order. This is true even when 1 and 2 can be construed as sequels to 3 and 4, i.e. as the goal of the movement in 3 and 4. One reason for this controlling function lies in the information density of pictures 1 and 2: whereas pictures 1-4 all portray space and person, only 1 and 2 include time and affective quality.

Another reason could be that viewers, in their interpretation of context, maintain a certain independence of meaning found in the individual picture signs, even when contextualization is not readily achieved (cf. the problem of whether a sequence depicts one or two separate individuals, as in 3.1 and 3.2). This means that the movement portrayals are terminated: they are used only in those contexts in which either the "destination/goal" (kommen) or the "departure-point/origin" (gehen) is assumed as "known from context-place" (Fillmore 1971:370). If this requirement cannot be found as a genuine property of the picture ordering, then the viewer himself ascribes it to the context. Interesting interpretations were also given for sequences 4.1 and 4.2:

1. The persons portrayed are seen as two separate people. Subjects who insisted that they "...knew Charlie Brown" and that the person depicted was one and the same idividuals were actually unable to interpret the sequence. The following statement is typical of their block: "I'm confused. All I can say is that there must be something wrong with the pictures."

2. Right-left or left-right rotation of pictures 1 and 2 has no effect on interpretation and cognitive content.

3. However, in order to maintain the dominance of pictures 1 and 2, subjects make assumptions about the relationship of the two persons and at times speculate on their emotional states, e.g.: "It looks like they've had an argument. The boy in the chair is relieved that the other one is going away." (for 4.2) or "It looks like they're not going to see each other for a while. The boy in the chair is sad that the other one is going away." (for 4.1)

The dominance of pictures 1 and 2 thus seems to be asserted by means of the emotional implications of the sequence. This occurs because

1. When judging sequences 4.1 and 4.2, all subjects interpreted the contents of picture 4 as source- or origin-oriented (e.g. Der geht aus dem Haus, 'He's leaving the house'; Er geht weg, 'He's going away').

2. Origin-oriented portrayals are context embedded and thus assume the context as given. If the sequence of the pictures is also understood as a temporal ordering of the content, then one possible description of 4.1 or 4.2 could be *The boy leaves the house and then sits down in the room. However, such a construct is semantically unacceptable. 4.1 and 4.2 are ungrammatical.

These investigations of the perception of left-right and right-left movement in picture sequences raise many questions. Moreover, in spite of the large sample, the results cannot be considered statistically valid. With the exception of the large proportion of left-handed participants, subjects were not selected on the basis of systematic criteria, and the study is, therefore, to be understood as a pretest to a more detailed investigation of the spatial code. The goal of this study has been to prepare hypotheses and to check the validity of the methodology, and results must, therefore, be viewed as preliminary.

Phase 1 Questions: The presentation of the pictures is rather an unusual artefact: in the most narrow case, pictures are bound within a frame: in the widest case, the physiological range of our field of vision can be considered such a boundary. In the present study, the edges of the paper might be regarded as a pseudoframe, although the arrangement of the pictures requires such direct focusing that the edges lie in the outermost portion of the field. However, this does not preclude a "wandering" eye movement as viewer's look for expected points of orientation. Does the assignment of kommen and gehen to right and left movement reflect a characteristic of the sign itself, or is this meaning ascribed by the viewer? Is an interpetation without context at all possible?

The subsequent investigation might show that the tendency to associate kommen with movement from left to right and gehen with movement from right to left is not consistent, but rather an equal number of subjects assign the direction right to left to kommen, and left to right to gehen. Would this mean that viewers are not restricted in the selection of perspective for context-free pictures?

Is the question of right and left a neurophysiological problem, as has been assumed in various attempts to explain space? If so, are interpretations predetermined?

Phase 2 Questions. Can it be demonstrated with statistical significance that portrayal of movement and spatial coding force the viewer to assume a certain point of orientation?

Does the orientation-point which a viewer establishes remain constant on the left, even when pictures are systematically varied?

When the figure moving within the frame is shifted, can we delineate a "switching point" at which either the picture becomes ungrammatical or the viewer's orientation point automatically changes?

Stage 3 Questions. Do the various codes yield a hierarchy that a picture with quantitatively higher value dominates a less valuable one when the two are interpreted?

Is temporal coding more dominant than spatial coding?

Does leftward or rightward rotation of a picture have a negative or positive affective influence? Is this influence generated on the basis of style elements which are to be placed on the sign level, or is it ascribed by the viewer on the basis of cultural norms and traditions?



3 similar properties apply not only to the perception of inanimate sequences of pictures, but also to the portrayal of movement in film and its perception by the viewer, provided that distinctions are made between the individual levels of film action.





Fillmore, Charles J. (1971) "How to Know whether You're Coming or Going", in Karl Hyldgaard-Jensen, ed., Linguistik 1971. Referate des 7. linguistischen Kolloquiums 11.-14. August 1971 in Kopenhagen. Athenäum, Frankfurt.

Fritsch, Vilmar (1964) Links und rechts in Wissenschaft und Leben, W. Kohlhammer, Stuttgart.

Mangan, James (1978) "Cultural Conventions of Pictorial Representation.

Iconic Literacy and Education", Educational Communication and Technology 26.

Piaget, Jean, Bärbel Inhelder, et al. (1975) Die Entwicklung des räumlichen Denkens beim Kinde, Klett, Stuttgart.

Stein, Nancy (1980) "Story Understanding", oral presentation at the Free University of Berlin.

Wertheimer, Max (1912) "Experimentelle Studien über das Sehen von Bewegungen", Zeitschrift für Psychologie 61.

Worth, Sol (undated) "Pictures Can't say Ain't," ms.

Worth, Sol and Larry Gross (1974) Symbolic Strategies", Journal of Communication.

Vernay, Henry (1974) Essai sur l'organisation de L'espace par divers systemes linguistiques; contribution a une linguistique de la tradtuction, Wilhelm Fink, München.






Bilyana Martinovski

University of Göteborg






1. Introduction

This paper consists of two main parts:

(i) a short analysis of the interactive functions of some deictic signs used in www in relation to written and spoken language;

(ii) discussion on the semiotic nature of deictic signs used in world-wide-web pages.

Let us start with the terms used in the title. The syntax of a human language is to a great extend deictic, it gives us an orientation in space and time with reference to the sender's and the receiver's positions. We distinguish between temporal, spatial, demonstrative and directional kinds of deixis (Lyons'77). Following Otto Jespersen (1922) R. Jakobson (1963) calls deictic categories for shifters in the sense that deixis is the link between the langue, the ritual, the code and the discourse, the actual, the message. The shifters are defined as indicial symbols because they represent conventionally and refer existentially to the utterance. In the following text I will observe that this definition is also applicable to the function of some deictic signs in www. Jakobson did not give his definition for the shifters in www, but for the shifters in the spoken and the written media. However, the new media examined here give new possibilities of referring to it's own reality.

The www-environment is quite visualized. The possibility of using images instead of linguistic expression tends not only to iconize but also to "indexalize" and "metaphorize" the interaction. In this regard it is interesting to see how the www-deictic expressions or tools are related to the written and spoken language deictic expressions. In order to do that I will examine the use of traffic signals.

2. Traffic signals1

One may distinguish the level of the narrative content from the level of narrative management. By narrative management is meant the strategies one makes use of in order to direct the readers or the listeners attention to actualized moments in the argumentation structure or the objective narrative content of the text. The traffic signals to be discussed below are part of the narrative management system. Examples of these are deictic expressions such as " I mentioned above...", "...later in the text...", "shortly", "further on", etc. Their main function is to guide the listener or the reader through the narrative content by providing information about the organization of a text, usually by means of anaphoric and cataphoric cross-references. These indexical devices typically involve spatial and temporal expressions like tense and deictic adverbs. In this context one may ask himself: do we conceptualize written and spoken language in the same way?

S. Fleischman’s answer (1991:295) is given in Table 1:

Discourse as time speech & writing earlier/


in a minute,



Discourse as space writing above


further on


Table 1

Written texts which are not supposed to be read infront of an audience, tend to use space referring expression, whereas spoken language and texts which are to be presented infront of an audience, like conference papers, use mainly time oriented traffic signals. The latter are also used in the written media.

From a diachronic point of view, it appears that the Guttenberg innovation has influenced the conceptualization of texts in a way which gives preference to the spatially oriented traffic signals.2 One may say that the visual reception is related to the spatial conceptualization of texts and the auditory reception is related to the temporal conceptualization of speech. This observation may be supported also by Jakobson's distinction between visual sign systems which he defines as spatial and audible sign systems which are temporal.

The internet media exploits both spatial and temporal traffic signals, that is, it exhibits both written and spoken language deictic features. To do that it uses linguistic expressions, pictures, numerals, colors, abstract signs. Instead of saying "here" or "this" this media uses an arrow, a hand, underlining, or color. Instead of saying "soon" it shows a clock which is blinking, and at the same time in another place on the screen - changing numbers. In Picture 1 below we have a clear example of the function of color change, namely, to show what is clickable, where can we find more information and to keep track of what has been already activated. The pointing hand, which usually has another form but turns to a hand when it is near (gets a contiguity relation to) a relevant text, is not a simple ostensive sign, it shows also that the pointed text contains information.

Picture 1

The indexical character of these signs which function also as whole expressions turns them into something more than traffic signals - it changes also their illocutionary force or gives the possibility to encapsulate in one sign more than one illocutionary acts. Pointing to something with a hand activates an information base and leads to another part of the text. It is not just a statement it is also a performance act. The main difference then between spoken/written language media and www media, with respect to the use of traffic signals, is the tendency to associate the www-discourse both with the concept of time and the concept of space by extended use of indexical signs, performing a number of acts simultaneously.

3. Signs and Relations

The second purpose of this paper is to examine the kind of "narrative management" signs used in www and to relate them to Peirce's semiotic system of signs and relations. I have put the term "narrative management" in quotation marks because, as discussed below, the internet media is not a narrative in the sense of a chronologically organized book or text. The following discussion will show that the internet media use many types of signs representing reality but the dominating type is the index.

3.1. The semiotic background

Let me first give a very brief description of Peirce's semiotic system.

Peirce's relations and signs





causality, contiguity, existential relation:



degenerated index



Table 2

A sign is something which stands to somebody for something in some respect or capacity. What it represents is it's object. The result of this representation is it's interpretant, which is also used to denote the interpreter. The sign represents it's object in a certain respect which is the ground of the relation between the sign and it's object. Thus the sign conveys information only on the basis of the relation to it's object on a certain ground. Consequently, types of signs are based on types of relations.

The signs are rarely of one type, the distinctions are thus theoretical. However, one of the relations may be dominating. Thus, for example, a photo-portrait is an icon, a footprint - an index, a word - a symbol. Short definitions of the signs would do for our purpose:

"A Symbol is a sign which refers to the Object that it denotes by virtue of a law, usually an association of general ideas, which operates to cause the Symbol to be interpreted as referring to that Object." (Peirce 1932:2.247)

"Anything whatever, be it quality, existent individual, or law, is an Icon of anything, in so far as it is like that thing and used as a sign of it." (Peirce1932:2.247)

What is most interesting for us is the character of the index. Peirce gives the following definition:

"An Index is a sign which refers to the Object that it denotes merely by virtue of being really affected by that Object." (Peirce 1932:2.247)

But there are also degenerated indices, that are anything which focuses attention (cf. Peirce 1932:2.265). Degenerated indices are not causal effects of objects, but devices which enable the interpreter to place himself in direct experiential or other connection with the thing meant.

Recognizing the impure character of the (second trichotomy of) signs Sebeok (1994:68) quotes Peirce: "it would be difficult if not impossible, to instance an absolutely pure index, or to find any sign absolutely devoid of the indexical quality" (1932:2.306), demonstratives and relative pronouns being the only potential exceptions (although they are linguistic signs, that is, symbols).

Sebeok (1994) discusses briefly two types of indices in Peirce: designators and reagents.

"Deictics of various sorts, including tenses, constitute perhaps the most clear-cut examples of designations" (Sebeok 1994:63).

An example of reagent would be "a piece of mould with a bullet-hole in it as a sign of a shot; for without a shot there would have been no hole, but there is a hole there, whether anybody has the sense to attribute it to a shot or not" (Peirce1932, 2:304).

Thus, the degenerated indexes are here called designators and the reagents are indexes. The last quotation suggests that the index (reagent) does not lose it's sign quality if there is no interpreting "mind", which is in a sense contradicting the initial definition of a sign (or my understanding of it). However, I will not speculate on this problem here.

There are number of other terms which have been associated with the concept of index.

"Such words as symptom, cue, clue, track, trail, and so forth, are among the high number of English quasi-synonyms of index." (Sebeok 1994:71)

We may now construct the following tentative delineation, which is a

mixture of Peirce's and interpretations of Peirce's definitions:





index (reagent)

degenerated index (design.)






causalitity (exist.)






























Table 3

3.2. The www-environment

Before relating the www traffic signals to the semiotic description from above let us first see what do we usually do when we connect to internet. As users of www we may have two basic types of behavior:

- we may search for something specific;

- we may "surf" or "stroll" in the electronic space without specific aim.

These search routines make use of different search tools. The above mentioned traffic signals, which in the internet media are often iconic signs, have a special function different from the function of the traffic signals in spoken and written discourse - they are predominantly search tools. The internet text is not physically organized as a book, it is not two-dimensional, but multi-dimensional or at least three-dimensional (like Escher's stairs which are connected in circles). Accordingly, as already mentioned, it exhibits features of a mixed spaciotemporal sign system. However, normally, not very experienced users tend to apply the image of a book, skimming through different pages. What kind of semiotic tools are we offered in order to search for specific information or just "surf"?

Following the definitions given above we may describe some of the www- signs and their types in the following manner:

language - symbols.

images - icons and indexes.

colors - index which may function as morphological endings having different meaning by convention or as anaphoric signs for already checked references.

underlining - symbol for possibility for further search or activation; its deictic character follows of its relation to other parts of the text which are not underlined and from the fact that it ostensively shows which parts are to be activated; thus it is also a degenerated indexical sign which we must pay attention to.

box organization - also a blend of symbol and index.

bold- style - functions as degenerated indexical sign.

flash - again degenerated indexical sign.

pointing hand - ostensive deictic sign, degenerated index, icon, symbol for "take".

arrow - spatial deixis, ostensive, icon, symbol.

clock - icon, temporal deixis, index, symbol (adverbial)

counter - numbers - symbol, index

Applying Table 3 on the www-signs and their functions I have constructed another table which describes in semiotic terms the character of the observed interaction managing www-signs.


traffic signal



index (reagent)

degenerated index (design.)





















bold style



































Table 4

Www obviously uses all kinds of blends of signs, but most of the traffic signals are indexes, both designators and reagents, that is, they are not only calling for attention but also "forcing the agent to accord them" (Peirce 1932:2.289). Unlike icons and symbols, indexicality is based on association by contiguity.

Indices, "whose relation to their object consists in a correspondence in fact [...] direct the attention to their objects by blind compulsion" (Peirce 1932:1.558).

The www-environment catches exactly this quality of the indicess, because the interaction proceeds quickly should and because of the immense amount of information in it. When searching something specific in the net we need management tools (traffic signal) which are easy to understand and which help us to deduct where to find the seeked information. In reality, however, we are often lost wondering which interpretation, which alternative of possibilities is to be chosen and in these cases we use reasoning to the best explanation (abductive inference), which does not give us any guaranties that we are on the right track. In the case of traffic signals this character of the interaction is also visualized and metaphorized (with associations to Sherloch Holmes’ kind of experience) by the use of magnifying glasses, keys, footprints, pointing hands, changes of colors, arrows, as well as much more elaborated pictures, as illustrated on Picture 2:


Picture 2

Here we have a different menu picture which relies on metaphoric interpretation of icons with indexical function. The concept of traffic signals is used both literally and metaphorically. Each of the images is clickable, that is, it is also an index, although this fact is not clear for a newcomer. The pointing hand helps us to understand what may be activated. We can activate the traffic signs - the car, the side signs, the post box, the sky, depicting the whirling information, etc.. If we have www-experience, we can guess what kind of information is hidden there, but not for all images. For example, we can not know what kind of information we can get from the "sky" or by clicking on the computer-building. We have to infer deductively and abductively a great deal of information.The indexical signs denote by virtue of being really affected by that object, the reality here being the cyberworld, manifesting its principles by these signs which the users try to decipher. Indexicalization triggers deduction, which consists of inferring important conclusions from seemingly insignificant clues (including our traffic signals which are associated with indexical topographical signs) in the process of locating information.

4. Conclusion

After examining www's interaction-managing signs we may now list the following characteristics of the www's interactive sign system:

- use of mixed spoken and written language features;

- mixed spacio-temporal conceptualization of the www-text;

- most traffic signals and procedures are indexical, trying to attract attention by blind compulsion and forcing the agent to accord them;

- the preferred representation of and interaction in the cyberreality is based on existential relations of contiguity and causality often in combination with resemblance;

- visualization metaphorizes the interpretation;

- the usual metaphor used in choosing icons for traffic signal is the detective metaphor.

The www interactive procedures are basically indexical, relying on abductive and deductive reasoning, based on existential relation of contiguity and causality. The new media exhibit both written and spoken language features, which, together with its visualization, define it as a mixed spatio-temporal sign system. It also gives excellent possibilities for metaphorization of discourse.



Jakobson, R. (1963) Essais de linguisticque générale. Paris: Editions de Minuit.

Jakobson, R. (1976) Språket i relation till andra kommunikationssystem. In:Tecken och Tydning,Till konstens semiotik. Texts selected by K. Aspelin & B. Lundberg, Stockholm: Pan/Nordstedt.

Jespersen, O.(1922) Language, Its Nature, Development, and Origin.. New York: Northon.

Fleischman, S.(1991) Discourse as space/Discourse as time: Reflections on the metalanguage of spoken and written discourse. Elsevier Science Publishers B.V., 1991.

Peirce, Ch. S. (1932) Collected Papers, ed. by Ch. Hartshorne & P. Weiss,

Vol. II,vElements of Logic, 2.228-2.272.

Perret, M. (1982) De l'espace romanesque · la matérialité du livre. L'éspace énonciatif des premiers romans en prose. Poétique 50: 173-182.

Sebeok, T.A.(1994) An Introduction to Semiotics. London: Pinter.



Kerstin Nelfelt

University of Göteborg






The Swedish sign language (hereafter SSL) is the language used in the deaf community in Sweden It is a natural language which has developed among the Swedish deaf. The social environments in which SSL develops are the special schools for the deaf and the societies of the deaf. A kernel of users, whose signing is seen as the model for SSL, are deaf persons who themselves have deaf parents. Different constructed signed languages and adaptations of SSL exists. However, they are not the object of this paper.

SSL, as well as other sign languages, are not universal and not mutually understandable. They differ especially with regard to lexicon but also with regard to morphology and syntax. There are, of course, factors relating to restrictions of perception and possibilities of articulation which are common to signed languages, as opposed to spoken languages.


Sign language research

In 1960 William C. Stokoe published a book called "Sign language structure: an outline of the communication system of the American deaf" (Stokoe 1960). This book can be said to be the beginning of modern sign language research. It is significant that the title uses the term communication system and not the term language. It had long been generally accepted that sign languages were not proper languages, i.e. did not fulfil the criterion used to distinguish between human languages and other types of communication. The distinction of Peirce between iconic, indexical and arbitrary symbols has, in linguistics, often been interpreted to mean that only arbitrary symbols are conventional symbols and thus linguistic. (e.g. Allwood and Andersson 1979). One of the arguments against sign language was that it seemed iconic in nature and thus non symbolic in the linguistic sense. For a long time, therefore, sign language research has been seen as an important goal to prove that sign languages are indeed languages and not some form of more primitive communication. One means has been to deny or at least diminish the importance of iconicity in sign language (e.g. Klima and Bellugi 1978, Ahlgren 1980). Bergman as early as 1979 (Bergman 1979) points out that the visual perception of sign language can be a basis for iconicity and that the combination of arbitrariness and conventionality, which has been accepted, without much reflection, as necessary and typical for human language, might be a coincidence, i.e. an effect of the auditive perception of spoken language.


The potential for motivation in visual systems

In contrast with dogs for example for whom smell is more important, or bats for whom hearing is most important, vision is probably the most important sense for human beings and our idea of the world is basically visual. However, spoken languages are based upon hearing not upon vision. This complicates the semiotics of spoken languages in a way not necessary for signed languages. Signed languages are based upon vision and this makes them closer to our visual concepts of the world and provides a basis for exploring semiotic motivation at different linguistic levels.

This is a similarity with writing and we know that the earliest attempts of graphic communication were iconic (e.g. Harris 1986, Lindqvist 1989). When not forced to adapt to auditive perception, people drew pictures of what they wanted to denote. Alphabetic writing was rather a result of the desideratum for more agreement between speech and writing.

Natural sign languages are not connected with, or dependent upon, language in the spoken medium and therefore, when they develop, can freely explore and refine the possibilities of visually based motivation.


The means of sign language

All expressive means in signed language must be visually perceptive. The means of SSL are manual gestures, mimetic/facial gestures, and body posture.

Manual gestures are gestures performed by the hands. The hands can be held in different configurations and in different positions and be moved in different ways and directions. The sets of forms, positions and movements being part of a sign language, e.g. SSL, are conventionalised in that language. Different sign languages use different sets. In hand configuration, movement and positioning there is a great potential for similarity to forms, movements and positions in the real world. This applies also to non verbal gestures. This potential can be exploited in word formation in particular, as well as in morphology and syntax.

A very special type of manual gesture is pointing. Pointing is a complex phenomenon in sign language and is exploited in sign formation, morphology and syntax. It can be made manually by the index finger, whole hand and other hand configurations, mimetically by gaze and also by body orientation.

Mimetics includes mimetic configurations and gaze behaviour. In spoken interaction these, with the exception of oral articulation, belong to non verbal communication and we are not very conscious about them (cf. e.g. Allwood 1976). In SSL, and also in other sign languages, these behaviours are used verbally and on a conscious level similar to that of other verbal behaviours. The meaning of instinctive body posture or orientation and facial expressions, e.g. surprise or puzzlement, are used intentionally for word formation as well as for grammatical purposes.


Word formation

The signs of the Swedish Sign Language are conventional symbols. The language has a lexicon of signs which are the conventional signs of the Swedish Sign Language, as opposed, on the one hand, to the gestures of non verbal communication and, on the other hand, to the signs of other sign languages. In SSL there is probably a majority of signs which are arbitrary with regard to form, i.e. there is no connection between the form of the signs and that which they denote (Bergman 1976). However, there are also very many signs where the form of the sign is motivated in some way by properties of that denoted.

I will describe different ways in which motivation as the relation between signs and denoted is exploited in word formation. In accordance with Lyons (Lyons 1977). I will not make semiotic distinctions within the set of motivated signs but rather describe different types of motivated articulation.

Single signs consist of manual gestures where the hand/s have a particular configuration (flat hand, claw hand, point hand etc.), perform/s particular movements in particular localisations in space. A specific mimetic expression, e.g. a certain mouth configuration can also be part of a sign.

Forms and contours can be depicted with the hands, either in configuration or by movements. Some salient property of that denoted is chosen and depicted. This is used in signs like TABLE1 , in which the surface of the table is shown by two flat hands (D-hands) turned downwards and moved outwards away from each other. In BALL, the form is depicted by two slightly bent hands (Y-hand) forming a globe and in ELK, both hands formed as elk horns (spread hand) are placed like horns at the temples. In the latter example, both the hand configuration and the localisation of the hands are motivated by the elk's horns.

The same hand forms can also be used in signs where they are not motivated, e.g. TO-BE-ON-THE-SAFE-SIDE, which consists of D-hand lightly knocking the chin and the same localisation as in ELK is used for example in MEAT. In the formation of both motivated and non motivated signs, only the set of hand forms which are permitted in SSL are used. Hand forms which are not part of the inventory of SSL (but which might be used in other sign languages) cannot be part of SSL signs.

Also the way things are drawn or formed in the air is stylistic and formalised, rather than naturalistic, e.g. the movements of the hands in BALL are like two parentheses rather than naturalistic half circles. However, the motivation gives a potential for improvisation in the form of a more naturalistic performance. You can use this for stylistic and humorous effects in poetry and jokes.

Actions and movements can also be imitated, for example, the movements of your arms and hands when fishing or driving a car. Such imitation can be used as a sign. Again in doing so, one is restricted to holding one's hands in a permitted form and performing permitted movements. Here there seem to be a greater degree of freedom when it comes to showing exact movement by modulating the sign, rather than adding descriptions.

Certain face expression, for example wrinkled lips to denote SOUR, can be used as signs of their own or as parts of complex (hands+face) signs. Notice that such a facial expressions are not expressions of feelings which the signer experiences but symbolic expressions. A mimetic expression, wide eyes and open mouth, is part of several signs sharing a meaning component of surprise. The mimetic part is necessary also in mentioning, as opposed to use, of the respective signs.

This part of SSL has not been studied in any detail to date and we know little about the set of permissible mouth shapes. It is generally assumed, though, that they form a limited, conventionalised set, just like the hand configurations.

Also, signs for abstract concepts can be motivated. As motivation can be used some concrete property associated in some way with the abstract phenomenon.


SIBLING - the bent index fingers are intertwined with each other

HELP - one hand lifts the other hand

In the first example the close family relation is shown as a close relation between the fingers. In the second case the abstract concept of helping is shown as the concrete act of lifting up someone or something.

The device is similar to metaphor and a good term for this type of motivated sign formation might be metaphorical word formation.


Sign modulation

There are different ways of modulating of the meaning of signs in SSL. It can be done by change of hand form, trajectory or place. It can also be done by repetition of signs and by holding or freezing the movement of signs. SSL also have grammatical morphemes, e.g. HAPP which denotes perfective aspect. It follows the sign it modifies but is not part of that sign. More than one device can be applied simultaneously.

In such modulations, motivation can be explored and I will give below some examples. Again, it is necessary to caution that this is an area with many white spots on the map. SSL research is very young compared to research about spoken languages and much is still to be done.



Change of hand configuration

As discussed in connection with sign formation, the hand configuration of a sign can have a meaning of its own. The hand configuration can also have a potential meaning. A finger is always one finger and it is always a longish thin object. Two fingers are, of course, two longish objects. Even when such a potential meaning is not exploited in a sign it can be exploited in sign modulation.


WEEK is signed with a hand shape of thumb and index finger (measure-hand). If the hand shape instead is thumb and index finger + middle finger, the meaning is 'two weeks', if it is signed by index finger + middle finger + ring finger the meaning is 'three weeks', etc.

LOOK, in citation form is articulated by outstretched index finger and long finger (V-shape). When performed with all fingers spread out (spread hand) it means that several persons are looking.

The first type of modulation is based upon the fact that one, two, three, four, five fingers respectively, denote the numbers 1, 2, 3, 4, 5 and that type of modulation can be used with a number of signs where the meaning incorporates some notion of number as, for example FIRST FLOOR, SECOND FLOOR etc., or FIRST PLACE (in a competition), SECOND PLACE (in a competition) etc.

Modulation of hand shape is also an obligatory device to obtain agreement of form between object sign and process sign, e.g. the hand shape of the sign GIVE must be modified according to the form of the given object. There are sets of hand shapes for this purpose which correspond to classes of objects, e.g. long slim objects, flat objects, vertical objects, etc. Probably modulations of hand shape can be used for sets of signs to denote corresponding sets of meaning variations Which sets of objects, modifications and meaning components are involved still remains to be studied and described for SSL.

Variation of trajectory

The trajectory of a sign can be modulated according to spatial relations.



LOOK, which in citation form is performed by a forward movement of the V-hand (cf. example above) can be modified by changing the direction of the movement towards the looked at object or towards the location of the looked at object. (cf. Syntax below)

WEEK can be modified to NEXT-WEEK by moving the right hand forward when concluding the sign and into LAST-WEEK by moving the right hand backwards.

The first type of modification is general in that signs implying an agent and/or patient and including a trajectory must always be directed in accordance with certain spatial locations, if the agent, and/or patient have been assigned such locations (cf. syntax below), .

The second type of modification is likewise general and used with a class of signs with a meaning element of time. The space behind the signer generally denotes the past and the space in front denotes the future.

Little research has been carried out in this and we do not know exactly which and how many processes of this kind are productive in SSL.

Fast and slow reduplication

Bergman describes two types of reduplication. The types of meaning modification connected with them are reciprocity and durational or habitual aspect (Bergman 1982). Reduplication can also be used to denote that there are several of that which the sign denotes.


KISS = kiss someone

KISS by both hands = kiss each other

WALK = walk, walks, walked

WALK repeated = walk for a long time or intensely.

HAVE-A-SCHNAPPS ˜ have one glass of strong liquor

HAVE-A-SCHNAPPS repeated = drink excessively, be an alcoholic.

FLOWER = flower, a flower

FLOWERX2 = flowers


Initial stop and initial pause

Bergman also describes two types of modulations, initial stop and initial pause, as connected with respectively inhibitory and gradual aspect.

There are no modifications corresponding to tense or gender or definite indefinite form in SSL.


The important motivated mechanism in SSL syntax is spatial order. In SSL the temporal ordering of items is not the only order or the most important order. More important is the spatial arrangement of signs. Manual signs are performed in the space before the signer. They can be performed to the right or to the left of each other, above and below each other, closer or further away from the signer and the movement always starts in one place and ends in another

Spatial arrangements of this type are used to mark agent and patient and generally to mark relations.


GIRL-place1 GIVE- slim-vertical-object place1->place2 FLOWER x 2 BOY-place2 = The girl gives a bouquet of flowers to the boy.

GIRL-place1 GIVE- slim-vertical-object place2->place1 FLOWER x 2 BOY-place2 = The girl is given a bouquet of flowers by the boy.

In the above sentences the boy and the girl are assigned certain localisations in space, place1 and place 2, respectively. Who is the recipient and who is the giver is marked by the direction of the GIVE sign. Note also that the GIVE sign must have a hand form in agreement with the form of the stems which you hold on to when handling bouquets.

Other types of relations spatially marked are, for example relative place in space.


CHAIR-place1 LAMP-place-to-the-left-of-place1 = The lamp stands to the left of the chair.

In this example the signs are placed in the signing space in the same way as the objects denoted are placed in real space.

Spatial arrangement can also be used for abstract relations and abstract objects. In more or less the same way as we use concrete metaphors for abstract phenomena abstract issues are described in spatial terms (cf. word formation).


CLEVER GIRL-place-high BOY-place-low = The girl is cleverer than the boy

CLEVER GIRL-place-high BOY-place-low -> place-high = The boy became as clever as the girl.

The usefulness of this type of spatially based motivated syntactic devices is enormous and it is used in extremely sophisticated ways. As with the other motivated mechanisms in sign language this too is limited by conventional rules but this is not extensively studied as yet. Like the other types of motivated devices in SSL it can also be used in an improvised and naturalistic way for jokes, poetry and probably for specification.


Summary and discussion

It is clear that motivation is an important relation between the language and the world in SSL. Motivation is exploited by means of conventionalisation and in a systematic way. It is obvious that conflicts can arise between the demands of systematicity and motivation. If a certain modulation or relation of signs is part of a paradigm and this paradigmatic modulation will turn out to be contrary to facts in the world for some individual sign, will the paradigm or the motivation win?


PASS-ANOTHER-CAR - one hand, in the shape proper for representing a car moves by the other hand in the same shape.

According to the general pattern, the right hand is always the active hand in right handed signers. This means that the right hand should be moved passed the stationary left hand. In Sweden we drive on the right hand side of the road, though, and cars pass on the left. According to one native deaf signer with whom I discussed this problem, the system point of view wins. The right hand is the movable and passes the left hand on the right hand side in contradiction with how real cars behave. However, I have witnessed a close description of an actual situation where a car passed the narrators car in an especially dangerous fashion. This description was done in a naturalistic way with the left hand passing the right and moving exactly as the culpable driver.

There is obviously much left to be explored in SSL with regard to the possibilities of motivated signing. A plausible speculation is that there is a kernel of syntagmatic and paradigmatic rules, but that the potential for motivation can be explored in a more improvised and naturalistic way for different purposes, for example in poetry, jokes and lively narrative.


Ahlgren, I. 1980, Döva barns teckenspråk, Forskning om teckenspråk VII, Stockholms Universitet, Inst. för lingvistik.

Allwood, J & L-G. Andersson, 1976, Semantik, GULING 1, Göteborgs Universitet, Inst. för lingvistik.

Andersson, L. and K. Nelfelt. 1994, Kurs i teckenspråk GLS 01 CD, Göteborgs Universitet, Inst. för lingvistik.

Armstrong, D.F., Stokoe, W.C. and E.W. Sherman 1995, Gesture and the nature of language, Cambridge University Press.

Bergman, B. 1977, Tecknad Svenska, Utbildningsforskning nr 28. Liber Läromedel/utbildningsförlaget, Stockholm.

Bergman, B. 1979, Dövas teckenspråk - en inledning. Forskning om teckenspråk III, Stockholms Universitet, Inst. för lingvistik.

Bergman, B. 1982, Forskning om teckenspråk XI, Stockholms Universitet, Inst. för lingvistik.

Gee, J. P. 1993, Reflections on the Nature of ASL and the Development of ASL linguistics: a comment to Corina's article, in G. Coulter ed. Phonetics and Phonology, Current issues in ASL Phonology 3. Academic Press.

Harris, R. 1986, The Origin of Writing, Open Court, la Salle Illinois.

Klima, E.S. & Bellugi, U. 1979, The signs of Language.

Lindqvist, C. 1989, Tecknens rike, Bonniers, Stockholm.

Lyons, J. 1977, Semantics, Cambridge University Press.

Nelfelt, K. 1986, Referentiell användning av mimik i svenskt teckenspråk, Gothenburg Papers in Theoretical Linguistics S 19. Göteborgs Universitet, Inst. för lingvistik.




Christiane Pankow

University of Göteborg






0. Preliminary remarks

In all languages there are many words and expressions whose references rely entirely on the circumstances of the utterance like a special situation or context; they can only be understood if these circumstances are known. This phenomenon in language is called deixis. Deictic words as in German ’ich’, ’du’, ’hier’, ’da’, ’so’ etc. can refer to persons, things, space, time and facts without naming them explicitly. In spite of their ambiguity deictic words are understood when a communication situation or a context is given.

In order to understand the complexity of deictic expressions, one should imagine a typical communication situation like the following.

Let us imagine a typical communication situation in second language acquisition. A student is to give a book review in front of his peers starting his paper in the following way: "I’ve read this big book here and now I want to talk about it." An accidental listener of this sentence could decode the meaning without a problem but he does not really know what book (title and author) the student is going to talk about. This communication situation often is accompanied by additional indicators not restricted to language. The student holds the book in one hand and points at the book with the index finger of his other hand looking in the direction of the book, then showing between the index finger and the thumb ’how big’ the book is. We only know that ’I’ refers to the sender, but we do not know who the sender is, ’here’ and ’this’ refer to the book, but we know neither the title and the author nor the size of the book. It is therefore obvious that personal and demonstrative pronouns, adverbs of place and time require additional indicators to enable us to interpret a communication situation. For that reason deictic expressions are also called pragmatic signs (cf. Pankow 1995: 572).

Another example will make the complexity of those signs clearer. The following instructions are included in a packet of Japanese chopsticks for German users:

(1) So halten Sie die Stäbchen in der Hand!

It follows a simple drawing, which shows in three steps what is meant by SO.


One could also describe how to use chopsticks with words and with gestures in order to understand what is meant by SO. ’So’ refers to an activity which is to be done. But how this activity is to be carried out, can only understood by looking at the drawing, because the adverb SO itself has no meaning of its own. We can only understand the sentence "So halten Sie die Eßstäbchen in der Hand!" [How to hold the chopsticks in your hand.], when, at the same time, we show verbally and/or using gestures what we have in mind. In written language, gestures and facial expressions are not included; the meaning of SO can only be understood by the verbal context or by other signs like graphics and pictures.

1. The meaning of the German adverb SO

Now, let us concentrate on the meaning and the usage of SO as a modal demonstrative adverb. Foreign students learn the correct meaning of the adverb SO relatively late. That is true for students of second language acquisition with Swedish as their native language. The usage of SO especially in written German is quite complex and does not correspond in all cases with the Swedish usage of ’så’. Compared to other demonstrative adverbs, information about SO is quite inadequate. The information we get in dictionaries is indeed more explicit but more concentrated on categorial classification like the usage of SO as an adverb, a particle or as a conjunction. In representative German and Swedish dictionaries SO and SÅ are described in the following manner:

SO (Adv.) (meist betont) bezeichnet eine durch Kontext od. Situation näher bestimmte Art, Weise eines Vorgangs, Zustands o.ä. (Duden - Universalwörterbuch)

adv. på detta vis, i den utsträckning, under sådana omständigheter. (Norstedts svenska ordbok)

It is remarkable that the Swedish SÅ is described using a definition in which new deictic expressions are used. The German dictionary Duden refers to its deictic meaning mentioning the fact that SO can only be understood in a special context or situation. Nevertheless, one can say that lexicographers make it easy for themselves by describing deictic expressions in this way. To a certain extent it is understandable because their meaning is the index function itself and deictic expressions themselves do not mean anything. That is why it is so difficult to explain them in dictionaries.

2. SO-Deixis in Karl Bühler’s Theory of language

The discovery of the pointing field [Zeigfeld] in language goes back to Karl Bühler (Bühler 1934). As his starting point, Bühler refers to the fact that the lexicon of a language can be divided into symbol words [Symbolwörter] and pointing words [Zeigwörter]. Therefore, he describes his theory of language as a two-field theory. He uses the classical terminology and calls them deictic expressions or pointing signs ([Zeigzeichen], Bühler 1934: 372). Furthermore, it becomes clear that the demonstratives are a special class of linguistic expressions, that differ from other expressions in a typical manner (cf. Ehrich 1983).

Bühler referred early to the fact that the listener starts an orientation procedure when the speaker uses a deictic expression. Using a deictic expression we can say the speaker refers to a concrete space of reference. Bühler also points out that the space of reference is not only based on a perception by the senses (for instance seeing the pointing finger as the basic referring manner), but different deictic spaces arise from special deictic procedures.

"Phänomenologisch aber gilt der Satz, daß der Zeigefinger, das natürliche Werkzeug der demonstratio ad oculos zwar ersetzt wird durch andere Zeighilfen; [...] Doch kann die Hilfe, die er und seine Äquvalente leisten, niemals schlechterdings wegfallen und entbehrt werden; auch nicht in der Anaphora, dem merkwürdigsten und spezifisch sprachlichen Modus des Zeigens. Diese Einsicht ist der Angelpunkt unserer Lehre vom Zeigfeld der Sprache." (Bühler 1934: 80-81)

Bühler says here that the pointing finger is the natural instrument of the demonstratio ad oculos and it can be replaced by equivalent pointing aids. But without a pointing help the specific linguistic pointing procedure could not be fulfilled. Even the anaphora, according to Bühler the most peculiar linguistic pointing procedure, needs a pointing help.

Bühler’s analysis of deictic expressions primarily covers the person deixis as well as the time and place deixis. But these do not cover the whole scope of deictic expressions in a language. We can certainly find solutions for deictic words like ’ich’, ’du’, ’hier’, ’jetzt’ concerning their space of reference [Verweisraum]. The space of reference is the room deictic expressions refer to. Depending on the specific deictic expression there are different spaces of reference. The natural space of reference is, according to Ehlich, the so-called ’speak-time-space* [Sprech-Zeit-Raum] (Ehlich 1987: 285). The deictic expression turns the attention of the listener directly to the perceptible object, which is pointed out in the perception room. This is also the reason why non-verbal, gestural signs appear in parallel with the deictic expression or even without it, for instance pointing out myself with the pointing finger means ’I’, pointing out the communication partner means ’you’, making a circular movement with one hand and looking at all those present means ’we’ etc. We can say that the gestures describe the basic meaning of the deictic expression.

(3) A: Mach mal so!

B: Wie denn?

A: So! (Person A shows person B a gymnastics exercise. ’So’ does not need to be named at all.)

Karl Bühler describes the meaning of SO just the way we did in our example above. He calls it so-deixis and presents it the following:

"Wenn ich ad oculos demonstrierend so sage, wird der Hörer auf irgendein aus der Wahrnehmung abzulesendes Wie verwiesen; ich mache ihm z.B. einen Handgriff vor oder zeige, wie das Produkt meiner Tätigkeit ausfallen muß... Und dieses Wie-Zeigmoment geht keineswegs verloren im Modus der Anaphora; es ist zu wenig gesagt, wenn man die so-Funktion als Unterstreichung (Emphase) charkterisiert." (Bühler 1934: 314)

When somebody demonstrating ad oculos says SO, the listener somehow gets referred to a ’how’ by making a movement with one hand and showing an activity. Bühler called this act of communication the reference of ’how’ [Wie-Zeigmoment]. According to him this is true even for anaphora. Furthermore, Bühler continues that it is not enough to characterise the function of SO as a simple emphasis.

It can be necessary to comment on Bühler’s observations on SO. There are two things to say that are connected with each other. Firstly, Bühler uses the term anaphora for deictic expressions in language. He understands them as a special case of deixis that refers to other linguistic expressions in speech. (Later, the term anaphora has become popular in syntax again, for instance by Wasow (1979), Webber (1979) and Thrane (1980). But we are not dealing with syntactic functions of SO here.)

Bühler continues his observations as follows:

"Psychologisch betrachtet setzt jeder anaphorische Gebrauch der Zeigwörter das eine voraus, daß Sender und Empfänger den Redeabfluß als Ganzes vor sich haben, auf dessen Teile man zurückgreifen kann. Sender und Empfänger müssen also dies Ganze soweit präsent haben, daß ein Wandern möglich ist, vergleichbar dem Wandern des Blickes an einem optisch präsenten Gegenstand." (Bühler 1934: 122)

Psychologically, the anaphoric use of deictic words requires that the sender and the recipient have access to a text as a whole, says Bühler, so that one could compare it with the roaming of the eyes on an optically present object.

Furthermore, Bühler also introduced the subdivision of anaphoric and cataphoric reference [Rückverweis, Vorverweis] and emphasized that he stood in a classical tradition of linguistics, and therefore he took the optical text as the picture of the language. But Bühler also thought that it is possible to describe oral speech, "die akustische Erscheinungsform der Rede" (Bühler 1934: 122), with his terminology.

Our second comment concerns Bühler’s ’Wie-Zeigmoment’ (the referring moment of ’how’). The type of the reference is obviously very important for the use of SO.

2.1 SO in the perception space

When the speaker says SO, the listener perceives even other (non-verbal) signs. ’So’ functions as a verbal signal for an activity that describes what is meant by SO. But this concrete moment of speaking in the presence of the speaker and the listener is not always given. Taking into account other linguistic contexts, the real difficulties of describing deictic expressions only start at this point. We can not fully explain the deixis of SO by describing the concrete perceptive signals of pointing only. Therefore, it is quite logical to try to remove the space of reference from the simplicity of the concrete situation of speaking and to look for other possible spaces of reference with a more complex deictic use of SO. Those kinds of reference can be established in various linguistic contexts. The abstraction of the deictic expression SO which was already pointed out by Bühler aims at different pointing spaces. We believe that the analysis of different uses of SO can only be done by integrating various possible spaces of reference (Ehlich 1987: 292). We want to show this by demonstrating linguistic data from a literary text, the novel Kassandra by Christa Wolf. The complexity of the linguistic procedures of reference becomesquite obvious in a concrete linguistic data base, in this case a literary text.

For this reason the novel has been scanned into the computer to get an easy access to all concrete occurrences of SO in the text. It is quite interesting that the number of indices is unexpectedly high: the index number of SO is 217, the absolute number of words in the novel is 46239. The following picture is given by comparing SO with other relevant deictic words.



Occurrence of some deictic words in the novel Kassandra (words abs.:

46 239)







sie (sg., pl.)









2.2 SO in the text space

The following example is typical for quite a number of occurrences of SO in the Kassandra text:

(4) [Ich legte meine Hand in Ihren Nacken, bis sie schwieg und wir beide, von der Mauer neben dem skäischen Tor, die Sonne ins Meer tauchen sahn.] SO standen wir zum letztenmal beisammen...

SO refers the attention of the reader to a part of the text which is located before the occurrence of SO. An activity [Hand in den Nacken legen] in a situation (a place [Mauer neben dem skäischen Tor] and a time [Sonne ins Meer trauchen]) is represented by SO and at the same time signalised as finished. Applying the above used term perception space, we are now going to describe the use of SO in a text space. There are some similarities between the perception space and the text space; but the reference of ’how’ is realised purely by verbal signs. The perception does not really take place, it is only an imagination. Furthermore, one does not perform the reference at the same time, as it usually happens in the perception space, but SO refers anaphorically to a place in the text. The form of this ’place’ could be a word, a phrase, a sentence or a whole section. Generally it seems that the space of reference for SO is much more difficult to describe than for other deictic expressions. And it seems to depend on the crucial character of the reference of ’how’. Syntactically it is difficult to separate. SO refers mostly to a whole section in a text, it is less concrete than the reference of ’ich’ or ’du’ that always refer to a concrete person named earlier in the text.

The text space of SO can consist of very complex meanings. SO can refer to situations, facts, or activities described in a special section of the text.

Here we have another occurrence of SO, in which SO refers back to a section in the text.

(5) Panthoos war ein Beutestück des Vetters Lampos von dem [ERSTEN SCHIFF]1 - SO1 nannte man das Unternehmen im Palast, als ihm ein zweites und ein drittes gefolgt waren und [man endlich die Bennenung des Volkes, "Schiff nach Delphi", durch neutrale Namen aus dem Verkehr ziehen wollte.]2 SO2 legte es kurz und bündig Anchises aus,...

In this occurrence SO1 refers back to [ERSTEM SCHIFF], SO2 is a reference back to [NEUTRALE NAMEN AUS DEM VERKEHR ZIEHEN]. The deictic procedure is intensified, because SO is placed at the first position in the following sentence. The anaphoric reference of SO links the previous thought with the new one. Here SO functions in a similar way to other sentence connectives, for example ’deshalb’, ’darum’, ’dadurch’ etc. It seems here that the reference to ’how’ has not been the first meaning of SO, but the anaphoric and connective function is obviously dominant. There is another anaphoric connective function of SO:

(6) Ich weiß, wer Kybele ist! schrie ich die Mutter an. SO, sagte Hekabe.

The use of SO is typical for dialogues. The addressee uses SO as an affirmation of the phrase of the other speaker. In opposition to ’ja’ SO refers more strongly to the type of the activity.

In the following occurrence we find another typical deictic function of SO:

(7) Da griff Freund Andron ein, mit seiner frischen Stimme.

[Warum den nicht! hört ich ihn sagen. Ach Schwester, dacht ich, könntest du ihn hören, deinen hübschen Tunichtgut. Am Abend, wurde abgemacht, würde sich Polyxena zeigen, auf der Mauer, neben dem Skäischen Tor, ihrem künftigen Besitzer. Inständig bat ich Polyxena, sich nicht zu zeigen. Warum denn nicht, sagte sie, wie jener Andron. Sie hatte keinen Grund dagegen, doch reichte das! Wo war ihr Grund dafür! So liebst du dieses Vieh? Kriegst du auch das noch fertig! entfuhr es mir. Der Satz, den ich mir nicht verzeihe. Der mir die Schwester unerreichbar weit entrückte. Ich sah es gleich, so wurde ihr Gesichtsausdruck: entrückt. Ich, in Panik, griff nach ihren Händen, entschuldigte mich, redete wie von Sinnen auf sie ein. Vergebens. Abends vor Sonnenuntergang stand sie auf der Mauer, mit jenem neuen fernen Lächeln, und blickte auf Achill hinab. Der stierte. Beinahe tropfte ihm der Speichel. Da entblößte meine Schwester Polyxena langsam ihre Brust, dabei blickte sie - immer wie von weit - auf uns: ihren Geliebten, ihren Bruder, ihre Schwester. Ich, flehentlich, erwiderte diesen Blick. He, Hektor! brüllte von unten mit heisrer Stimme Achill das Vieh. Hörst du mich! Die Übereinkunft gilt.

Sie galt. Für Monate war meine Schwester Polyxena die bewundertste Frau in Troia. Das hatte sie gewollt. Die Ihren strafen, indem sie sich selbst verdarb: Die Taten, die der Krieg heraustrieb, waren Mißgeburten. Polyxena hatte, als sie ihre Brust dem Griechen hinhielt, das Kind des Andron als ein kleines Klüpchen Blut verloren. Triumphierend, schamlos gab sie es bekannt. Frei sei sie, frei. Nichts, niemand halte sie.]

SO war es.

SO refers here to a finished narrative unit concluded by the phrase ’So war es.’ The meaning of SO is reduced to its connective function (cf. Burkhardt 1987: 307); it keeps the narrative units together and it always refers to the previous narrative unit. Thus SO functions as an anaphoric expression at the same time producing a turning point of the narrative process. Its basic meaning, the reference of ’how’ is strongly faded.

2.3 SO in the imagination space

There is a large number of occurrences of SO that cannot be classified either to the perception space or to the text space. The question of the very nature of the deictic procedure arises again. How should the usage of SO in the following cases be interpreted?

(8) [...] Dies alles war SO unvorstellbar. SO dumm, SO schlecht erfunden, daß ich, ein Kind, nur bitten konnte, mich damit zu verschonen. Doch juckte das Thema SO sehr, machte mich SO scharf auf jeden Gesprächsfetzen,...

The problem gets still more complicated, if the meaning of SO increases by the repetition of the same symbolic sign.

(9) SO viele Brüder, SOviel Kummer. SO viele Schwestern, SOviel Entsetzen.

In these cases SO is not a text structure related instruction like in examples (4) to (7), though there is a deictic procedure. But what does this deixis mean? The answer is, that the deictic spaces have been changed. This deictic procedure takes place in an imagination space, and it differs clearly from all kinds of local deixis. Establishing the local deixis, the reader uses a wide variety of expressions from the symbolic field of the language helping him to point out the objects that the deictic procedure refers to (cf. Ehlich 1987: 297-298). This kind of instruction aid are missing in example (9). We can hardly recognise the quality of ’unvorstellbar’, ’dumm’, ’schlecht’, and ’viel’. How do we have to understand the deictic procedure in the imagination space?

The text producer - in our case Christa Wolf - makes great demands on the reader, because the reader himself must design this imagination space. The reader must be able to copy the poetic imagination of the text producer. How the reader - the recipient of the text - gets the same imagination object as the author, is not foreseeable.

According to our discussion of occurrences of SO in a literary text there should not be a difference to other types of text or communication situations like for instance expressions in everyday life.

(10) Er ist so lieb zu ihr.

The emphatic usage of SO in everyday life makes the same demands on the dialogue partner when it comes to its imagination space.

3. Conclusion

Let us summarise the results of our observations on the meaning of the deictic adverb SO in German. What SO means, can be described by deictic procedures that can be called orientation procedures. One can also say, when a speaker uses SO, the dialogue partner activates different kinds of orientation procedures, and as the result of it he starts focusing a space of reference. In opposition to other deictic expressions, SO can refer to activities and complex facts. Bühler calls this quality the reference of ’how’ [Wie-Zeigmoment]. The spaces of reference discussed above i.e. perception space, text space, and imagination space seem to be basic for the meaning of SO. It does not exclude the fact that there are other possible spaces of reference for describing the meaning of SO. But it is quite obvious that the semantic analysis of SO can be relatively well presented with the help of different kinds of spaces of reference.



I am grateful to Norbert Richard Wolf (Maximilian University of Würzburg) for using his KWIC-Index of the Kassandra text.



Burkhardt, A. (1986): SOSO? Kritik und weiterführende Überlegungen zu Konrad Ehlichs Aufsatz über die Funktionen des deutschen SO. In: Rosengren, I. (ed.): Sprache und Pragmantik. Lunder Symposium 1986. Stockholm. 299-314.

Bühler, K. (1934): Sprachtheorie. Die Darstellungsfunktion der Sprache. Jena.

Duden. Deutsches Universalwörterbuch (1989). Mannheim, Wien, Zürich.

Ehlich, K. (1986): so - Überlegungen zum Verhältnis sprachlicher Formen und sprachlichen Handelns, allgemein und an einem widerspenstigen Beispiel. In: Rosengren, I. (ed.): Sprache und Pragmantik. Lunder Symposium 1986. Stockholm. 279-298.

Ehrich, V. (1983): Da im System der lokalen Demonstrativadverbien des Deutschen. In: Zeitschrift für Sprachwissenschaft 2,2. 197-219.

Norstedts svenska ordbok (1990). Stockholm.

Pankow, A. & Pankow Chr. (1994): He said he would bloody come back tomorrow. Zur Verschiebung der Zeit- und Raumdeixis in indirekter und erlebter Rede. In: Festschrift für Astrid Stedje. Umeå Studies in the Humanities. 153-168.

Pankow, Chr. (1995): Semiotics. In: Handbook of Pragmatics. Eds.: Jef Verschueren, Jan-Ola Östman, Jan Blommaert. IPrA Research Center (IRC). John Benjamins. Amsterdam, New York. 569-476.

Thrane, T. (1980): Referencial-semantic analysis. Cambridge.

Wasow, Th. (1979): Anaphora in generative grammar. Gent.

Webber, B.L. (1979): A formal approach to discourse anaphora. New York.

Wolf, Ch. (1984): Kassandra. Vier Vorlesungen. Eine Erzählung. Berlin, Weimar.



Roland Posner

Technical University Berlin







"That wasn't Yeltzin, that was his double."

(Commentary by Jürgen Thebrath

in the German TV news

Tagesthemen of Feb. 16, 1995)




1. The role humans play in models of communication

"Culture communicates", says Edmund Leach (1976:2). To understand culture, therefore, one seems to need models of communication. Let us examine such models with respect to the role they attribute to humans.

Shannon and Weaver (1949) and Meyer-Eppler (1959) specify the factors which must be involved in a process so that it becomes a process of communication. It is necessary to have a sign (i. e., a signal) and a channel, through which it reaches a recipient, and in certain cases there is also a sender, from which the sign originates, as well as a code, which provides the signifiers and signifieds, according to which the sign must be interpreted if a message is to be drawn from it.

Various types of interaction are assumed to occur between these factors, yet the question, which of these factors can be embodied by humans, is answered in the same way by all communication models. There is a general tendency to define the communication process without recourse to humans. Shannon and Weaver (1949:3 and 33ff), for example, design a chain of communication, which connects the information source with its destination via a sender, a channel, and a receiver, but consists merely of a series of technical apparatuses linked with one another.1

A different picture arises when one scrutinizes the verbal descriptions formulated by the authors of these communication models; they are heavily based on anthropomorphic ideas (cf. Cherry 1957=1967:260): There is a sender, who has the intention of sending a message to a recipient. He chooses a code which he believes to be mastered by the recipient, and he selects from it the signifieds appropriate for the intended message. Since these signifieds are paired with certain signifiers through the code, the sender then produces signs which embody these signifiers. He sends them through the channel to the recipient, who goes through the same chain of factors in reversed order if he has the intention of drawing a message from the sign (cf. Posner 1989:244f).

As a result, most people who apply models of communication in the human and social sciences equate man's role with the roles of the sender or recipient in communication processes and do not consider whether humans might also be signs, channels, codes, signifiers, signifieds, or messages (cf., e.g., Jakobson 1960:357f). Saussure equips the sender and recipient with a brain, mouth, and ears, and describes the "model of a speaking-circuit" (1916:28 = 1964:11ff) as follows: "Suppose that the opening of the circuit is in A's brain [...]. A given concept unlocks a corresponding sound-image in the brain; this purely psychological phenomenon is followed in turn by a physiological process: the brain transmits an impulse corresponding to the image to the organs used in producing sounds. Then the sound waves travel from the mouth of A to the ear of B: a purely physical process. Next, the circuit continues in B, but the order is reversed: from the ear to the brain, the physiological transmission of the sound image; in the brain, the psychological association of the image with the corresponding concept."

In such a context, it could appear erroneous to speak of man as a sign. And it seems to fit this context well when Roland Barthes (1975:84) remarks:

"I live in a society of senders: each person I meet [...] sends me a book, a text, a result, a prospect, a petition, an invitation to an event or an exhibition, etc. The pleasure to write, to produce, presses from all sides."2

2. Humans as images of god

There were times in which humans in the West did not consider themselves primarily as senders who produce signs, but at best as recipients who must interpret one another as signs. At the beginning of the 13th century the German poet Freidank wrote, "Nehein geschepfede is sô frî, sin bezeichne anders dan si sî" ("No creature can avoid denoting something different from itself"; Freidank's Bescheidenheit 12, 11-12). The justification

for this claim is to be found in the Bible (Genesis I, 26): "Then God said, 'Let us make man in our image, after our likeness' [...] . So God created man in his own image, in the image of God he created him; male and female he created them" (English translation according to the Revised Standard Version).

As image of God every person is a sign, and the Creator must be conceived as a sender according to the communication models. The God of the Jews and the Christians created Adam from earth and Eve from one of Adam's ribs so that they might refer to Him as their Creator. Thus, every human can be understood as God's means of addressing the other humans. Those who regard a person as a sign of God in this way may find it difficult to accept that the very person might simultaneously be a sender in the sense of the communication models: Whenever a person speaks, they will assume that someone else is speaking through him -- either God or the evil spirits who instrumentalize him. In the Western tradition, God's medium of addressing people through the body of a given person is that person's soul or heart.

How can a communication theorist in a "society of senders" work with this view? He interprets the Bible in a new way by maintaining that it was not God who created man in His image but rather that it was man who imagined God to be of his own kind and who invented the myth of the Creation solely as legitimization of this theological idea.3 Formulations which point to such an interpretation can be found even in Luther; in his commentary on the first Commandment, "Thou shalt have no other Gods before me", in the Great Catechism of 1529 one can read: " [...] trust and faith alone maketh both God and idols" (Werke 30, 1:132,32-133,8). Later commentators (e.g., Ebeling 1964:42) are certainly correct in emphasizing: "It would be foolishness to interpret [Luther's] [...] phrase to mean that man is a creator and God a mere creation." Yet exactly this interpretation was put forth explicitly by Ludwig Feuerbach (1841), who viewed God as a product and projection of mankind. This thought persists in modern theological considerations (cf., e.g., Karl Barth 1928).4

The inversion of the idea of the creation posits man as a sender, and this sender not only creates signs existing independently of him, such as images or texts (Aleida Assmann 1996: § 1, calls these "excarnated signs"), but he also can posit himself as a sign when necessary (i.e., he can appear as an "incarnated sign"). Roland Barthes, who will serve us as a source for contemporary views in the following, pronounces the latter intention in this way (1975:145): "I had only one solution: to write myself anew -- to add another kind of statement to the books, topics, memories, and texts."

This "other kind of statement" does not consist in adding to the already published works further sign-vehicles, which then have an existence external to the sender, it rather consists in using one's own body as a sign-vehicle, as animals do. "Hardly touching it, I cover the written work, the bygone body and corpus with a kind of patch-work, a rhapsodic quilt composed of squares sewn together." In order to discover which message man as a sign can convey after the exclusion of God from the semiotic discourse, one need no longer drive to the core of man's being and involve his soul or heart: "Far from immersing in myself I remain on the surface, because this time it's about 'me' (about my ego) and because the depths belong to the others" (Barthes 1975 145).5

He who presents himself as a means of communication in Barthesian fashion will seldom be satisfied with his surface as it is. Taking over God's work, everyone creates oneself anew "as a man" or "as a woman". As early as the Stone Age, tattooing was supposed to support this semiotization of the body.

The man who froze to death on a Tyrolean glacier 4600 years ago had a tattoo on his body, which stands open to interpretation once more today. In the Palau Islands, as late as the 19th century, "the more tattoos a woman had on her pubic region, the more desired she was" (Probst 1992:22).

No wonder that God the Creator, in the Jewish or Christian sense, sought to prevent such interference with his creation: "You shall not make any cuttings in your flesh [...] or tattoo any marks upon you: I am the LORD" (Leviticus XIX, 28, Revised Standard Version). He who changes a sign conceived to be an image produces misunderstandings about its referent.

European women and men use much less incisive means of self presentation today: she applies make-up, mascara and lipstick, decorates herself with earrings, necklaces, and bracelets, or wears blouses, bodystockings, and stiletto-heeled shoes. He applies a razor and beard-trimmer, puts on hornrimmed glasses, and wears a shirt and tie. Both put on trousers and blazer, cap and scarf, if they don't leave it at sweaters and jeans.

3. Humans as utterances

Both the anthropomorphic communication models and the secularized idea of creation impose limitations upon the roles man may play in the sign process. They insinuate that a person is either only a sender or only a sign. If man is a sender, it seems impossible for him simultaneously to be a sign, because he would then not exist until he had been produced as a sign. If he is a sign then it seems impossible for him to function as a sender, because as a sign he must have been produced by another sender and intended for (yet another) recipient.

Such paradoxes become especially noticeable if the sign is conceptualized as an utterance: as speech directed at another person in a process of communication. Roland Barthes loves places in which the roles of a sender, sign, and recipient, imposed upon the human body in the framework of communication theory, become indistinguishable, for example, a "Bar in Tangiers" (Barthes 1975:144f): In contrast to normal establishments of this type where "there is speaking, communicating, meeting", this bar is "a place of semiabsence. The room is not without bodies; they are even quite close, and that is the important thing" -- yet they remain anonymous.

"In this bar an other's body never becomes a 'person' (with its personal data [...] etc.): I am invited for a walk, but there is no address directed at me. [...] The bar is [...] a neutral place: the utopia of the third concept, a drifting far away from the too-pure pair speech /silence ."6

Nevertheless it is interesting to see what happens when people get involved with this division of roles and try to embody several roles simultaneously (Barthes 1875:144):

"[...] speaking while kissing, kissing while speaking. One should suppose that there is this desire since lovers constantly 'drink words from lovers' lips'. What they are enjoying there is [...] the stammered body."7

With these formulations, Barthes makes clear in succinct brevity what is characteristic for loving. Lovers are bodies which enjoy (thus recipients) as well as bodies which are enjoyed (thus signs) as well as bodies which control the enjoyment of others by their behavior (thus senders). They present themselves as utterances directed at one another.

Other kinds of people's behavior can be understood as utterances as well. As an example, consider that which we all do every morning. Washing oneself, drying, putting on make-up, clothes, and jewelry: isn't it like a process of self-formulation in slow-motion? That which strides through the partment door, goes to the newsstand, enters the subway, rides the elevator and sits down in the office is no longer a "stammered body" in Barthes' sense but a well-articulated sign: the preened image of a business manager, the carefully made-up face of the secretary, the consciously presented body of the model, the sloppily formulated audacity of the T-Shirt-wearer with blue-jeans and sandals, the hastily prepared appearance of the schoolgirl. All these are signs facing our neighbors, who are signs themselves, but feel addressed as recipients and inclined to respond as senders.

The analogy between verbal utterances and bodily demeanor can be pursued further into the modalities of communication. As in verbal utterances, one can distinguish between situations of production and reception of signs in daily interaction. Clothing, make-up, and jewelry cannot be adorned and set aside as quickly as one can articulate and correct oral utterances. Grooming as a situation of sign-production corresponds rather to sitting down at a desk and formulating a written manuscript for a speech which is to be read at a later date, when it will be correctable only with difficulty. The paperboy, the crowd in the subway, persons in the elevator, the colleagues at work react to an image that was created hours before -- not to mention the fact that the clothes can become disheveled, the jewelry lost, and the make-up smeared in the rain.

As in verbal utterances, the interpreter of bodily demeanor can never be certain that he draws the line between information conveyed intentionally and unintended conclusions properly. An important criterion for differentiation is the changeability of the bodily and behavioral features in question. Signs that are easily changed are more often suspected to be manipulated than others: "She proclaimed her innocence, but did you notice the way she was sitting there?" "He presented a perfect show as a mayor with the chain of office around his neck and the signet ring on his finger, but his hands moved like those of a peasant." As remarks of that kind demonstrate, we examine the gestures and posture in order to ascertain that we may trust the oral and facial expressions of a person.

To be certain that we can trust the bodily movements and clothes we test the more permanent characteristics of the body. This creates a hierarchy of credibility which relates life-long characteristics (fingerprints, voice patterns, eye color, primary sex characteristics, facial structure, skin color, birthmarks) to those lasting many years (body size, scars, mutilations, prosthetic devices), those lasting several days (hairstyles, type of beard, length of nails), those remaining the same for several hours (clothes, jewelry, make-up), and those lasting only minutes (bodily position and posture) or changing by the second (facial expressions and gestures), or fractions of a second (speaking, laughing, sighing). This hierarchy is only interrupted by bodily reflexes (blinking at light disturbances, hiccoughing, sneezing), which last only seconds and nevertheless count as behavior that is not manipulated.

Every human utterance is subject to this hierarchy of credibility and is almost automatically submitted to a contextual comparison. Therefore, persons who question the validity of their permanent characteristics (facial structure, primary sex characteristics, hairline, etc.) through surgical alterations (see Moser 1988) put themselves in a difficult situation. If one's long-term image appears to be manipulated, the credibility of one's shorter utterances is also weakened.8

4. Humans as signifiers

The hierarchy of credibility demonstrates that not all of a person's characteristics can be classified in the same way as utterances. Personal utterances are generally considered as intended, even if some of their aspects are not under full control, and they are typically short-lasting phenomena, even if they can achieve permanence through recording procedures. This justifies regarding not only linguistic formulations but also facial expression, gesture and posture, clothing, jewelry and make-up, length of fingernails, and hair- and beard styles as utterances produced by a person for the purpose of self-presentation.

Yet the utterance analogy fails when it comes to body size and pupil dilation, skin color and fingerprints. These signs are either very difficult to control intentionally (body growth and pupil dilation) or hardly modifiable (pigmentation distribution in the skin and the pattern of ridges in the fingertips). Therefore they cannot be produced as signs even if they are received as such. They are interpretable only because they belong to a field of signifiers (Prieto 1966:37=1972:39f: "sematic field") which is structured in correspondence to a field of signifieds (Prieto 1966:35 = 1972:39: "noetic field").

Thus, before colonialization, people's skin color was seen as strongly correlating with their geographical place of origin. Even today, Africans from further south than the Sahara are referred to as "black", Asians as "yellow", Indians as "red", and Europeans as "white", and the consequences, attempts at mutual spatial exclusion and dominance, are well-known.

An African who paints himself white does not thereby become a European, and a European who paints himself black is still a European. They are just as incapable of changing their skin color as their roots, and for this reason the skin color can signify the origin of a person even though that person cannot use his skin color to communicate anything. However, an African who puts a white make-up on his face and a European who colors his skin black can communicate through this make-up. If the context is appropriate, they will tell their companions: "Imagine I were white and came from Europe!" or "Imagine I were black and came from Africa!" These are utterances in the sense discussed above.

If one regards humans as signs, one must therefore differentiate between communicative signs produced by a sender with a particular communicative intention, and significative signs. The latter embody particular signifiers which refer the recipients to corresponding signifieds if they are familiar with the appropriate code, even if they were not produced for this purpose.

Signification is a sign process which can take place without anyone intending to send a message to someone. Through signification, a knowledgeable interpreter can be informed about a person without that person desiring so or ever having thought about it. The codes that underly such signification processes have been investigated since ancient times by medical semiotics (cf. Hucklenbroich 1996: § 2.3). The resulting knowledge can engender great advantages for the sign recipients indeed, often at the expense of the unwitting sign vehicle. If an innkeeper recognizes the worn wrinkles around the mouth of his guest as a sign of stomach troubles, he will serve him differently than a gourmand. If a saleslady notices the dilation of her customer's pupils occurring when he looks at a particular item, she will not comply even with the most stubborn haggling, because she knows that he will purchase the item anyway. A person who raises his or her eyebrows can communicate interest; a person whose pupils dilate, however, does not communicate anything, since he cannot prevent his body from signifying his interest.

Signification is possible without senders, but it can also be utilized by persons who make themselves into signs. Elaborate systems of signification (i.e., codes) have been developed for this purpose in all societies. Clothing is a central example. Monks consider themselves as living symbols of God (Li 1996: § 9.3), but they can only be recognized as such by their robes. Police officers are regarded as walking signs of state authority, but without their uniforms they are hardly identifiable. The clothing someone wears often serves as a sign of his profession. And this is put to use by larger organizations who design their own clothes and differentiate them according to the various fields of activity of their members. In the armed forces of most countries, the marines have a uniform different from that of the army and the air force. And each unit uses epaulettes as signifiers of rank. (Civil armies like those of the French or Chinese revolutions show little differentiation, while professional armies carry the differentiation to extremes.

The codes in use also make it possible for people to utilize other people as communicative signs. Many an industrialist who marries an actress does so to present her as a signifier for his appreciation of art. Some European families who adopt an orphan from the Third World do so to create a signifier for their sense of responsibility. He who surrounds himself with many servants often uses these as signifiers for his wealth. All status symbols are communicative signs based on signification, and, for this reason, no one is safe from being misused as a status symbol.

Charles S. Peirce also analyzed man as a signifier. He does not seek man's sign properties on the surface, as does Barthes, but within. In the style of the closing 19th century, he takes thinking to be the basic semiotic activity of man, not communicating (cf. Posner 1995).

According to Peirce, every thought we have is a sign of its content, and by every state of thought we refer to the object thought of. Since a living person is unable to avoid thinking, one's life can be seen as a chain of thought-states continuously replacing one another. And if every state of thought is a sign, then the complete chain of thought-states is a complex sign:

"[...] the fact that every thought is a sign, taken in conjunction with the fact that life is a train of thought, proves that man is a sign" (Peirce 1931-58: § 5.313).

It is interesting that Peirce takes into consideration the development of each person which leads to progress in his thought. In doing this, however, he reformulates the question slightly. Regarding the linguistic expression as prototypical sign in his essay "’Some consequences of four incapacities", he asks: "What distinguishes a man from a word?" (Peirce 1931-58: § 5.313).

According to Peirce, there is a whole range of differences:

-- The signifying characteristics and the forces in man which determine what he means and to what he refers are all exceedingly complicated compared to those of the word. Yet Peirce takes this only as a relative, not an essential difference.

-- Man has consciousness, words do not. But consciousness is for Peirce only a certain kind of self-perception, and thus this difference also belongs to the signifying characteristics of a person and remains external to him.

-- Man as a sign is able to take in information; his meaning becomes enriched; he can mean more and more. But this is also true of words, says Peirce: "Does not electricity mean more now than it did in the days of Franklin?"

Having rejected these attempts at differentiation, Peirce emphasizes: "Man makes the word, and the word means nothing which the man has not made it mean." This amounts to assuming a two-level structure for the human sign: Man is himself a sign, and he creates signs; he is a sign-creating sign.

Yet Peirce makes it very clear in this context that a gain in information would be impossible for man if he could not use words to carry this information as their meaning. Man can only be a sign by creating signs.9

It is remarkable how Peirce takes it for granted that man has a status comparable to that of a word. The biblical conception is still at work here: Christ as incarnation of God's word and humans as creatures who strive to attain this status in the imitation of Christ.

Peirce goes so far as to say: " [...] the word or sign which man uses is man himself" (1931-58: § 5.314). For him, the consequence is then: "My language is the sum total of myself" (1931-58: § 5.314).

These conclusions reached by Peirce in the year 1868 are reminiscent of Wittgenstein's claim of 1921: "The boundaries of my language are the boundaries of my world" (Wittgenstein 1922: § 5.6). Both philosophers consider language to be a system of verbal signs, but Wittgenstein refers to their signifieds and Peirce to their signifiers. Neither of them is concerned with the individual linguistic utterances made by humans, but with utterance types, that is, words and sentences as elements of a language system.

5. Humans as actors

From the means of writing oneself let us now revert to the proper acts of writing, to the uttering processes. If a person were nothing but a system of signs, as Peirce and the early Wittgenstein assumed, then his utterances would merely be components of this system, realizations of ready-made signifiers. But is this always the case? A glance at the theater teaches us otherwise.

Every actor utters words and sentences in order to be taken not as himself but as another person. Is he not expressing himself in these utterances? Does he not act as their sender?

In answering these questions we might find it instructive to consider how we are used to speaking about expression. We say, "Someone expresses an opinion" and "He expresses himself", "He expresses himself about something else" and even "He expresses himself about himself'. But not everyone who expresses an opinion about something expresses himself with regard to it; he can speak as another person's mouthpiece. Persons who express opinions about something without expressing themselves with regard to it are said to play roles.

The question whether someone is expressing himself or just playing a role can be answered on the basis of discrepancies between his utterances on the different levels of the hierarchy of credibility. If such discrepancies occur, the less permanent utterance is taken to be improper or fictional. This facilitates acting in theater.

In everyday life, a person's behavior on the different levels of the hierarchy of credibility serves the others to identify that person. Each person has a set of personal characteristics on every level of the hierarchy; they constitute the specific personal code. But in most situations only a small selection of the personal code is utilized in identifying a person. A fleeting glance at the wellknown sweater or half an ear's listening to the person's voice is usually sufficient. One looks or listens twice only when discrepancies appear (unusual clothing, a sniffly voice, an unfamiliar way of walking). These are the cases when longer-lasting characteristics such as facial structure or physique are taken into account.

In the theater, this selective method of perception is systematically exploited. It is only by partial identity that an object on stage refers to the object presented. A piece of furniture on stage stands for a piece of furniture of the same size and form (but possibly of a completely different material); the light on stage stands for light of the same brightness and color (but

possibly from a completely different source); the person on stage stands for a person with the same clothing and hair-style (but possibly someone with a completely different language and knowledge, feelings and attitudes). A comparatively small complex of prominent characteristics shared by another person is used to refer to that person and to call to mind all the rest of his permanent characteristics.

Thus for the audience the actor is on the one hand an iconic sign with behavorial characteristics shared by another person, and on the other hand an indexical sign referring to that person's other characteristics which he does not share (cf. Rozik 1995). This is also valid in film. And if the person presented has only a fictional existence, then his characteristics not shown by the actor can remain indeterminate (cf. Ingarden 1931:261ff=1973:335ff).

These structures also allow for the presentation of emotions on stage (cf. Ekman & Friesen 1975). Any actor who wants to display surprise must be able to produce some of the facial traits (or gestures) which the person presented would have in a moment where he would experience surprise. The presentation of emotions is thus achieved through imitation of prominent parts of actual affective behavior. The display of such facial traits (or gestures) by the actor creates a reference to the corresponding emotions, and the prominence of such facial traits (or gestures) in the corresponding emotional behavior makes the audience supplement the other characteristics of the face (or gestures) which belong to that emotion.

This is, again, a two-phase sign process based on iconic reference and its indexical supplementation.

With the proper choice of iconic and indexical reference, acting can itself be portrayed. This requires, on the one hand, the production of prominent characteristics of the person that is presented by the actor to be portrayed, and, on the other hand, the production of prominent characteristics of the actor to be portrayed which are missing in the person presented. Combinations of contradictory characteristics are particularly effective, such as the linking of flexible singing with stiff bodily motion in the portrayal of an opera singer who is to play a seducer, or the linking of foreign names with English pronunciation in the presentation of a foreign play on an English stage.

As this example demonstrates, the audience in a theater not only realizes iconic references to prominent characteristics of the person presented and their indexical supplementation, but also pays attention to prominent characteristics of the actor's behavior which constitutes the core of his own personality. In this way, actors become two-level signs: By expressing something on behalf of another person, they are able to express themselves.

This raises the question of how the actor's behavior differs from the role playing of people outside the realm of the performing arts, e.g., politicians. The answer is simple: An actor must noticeably distance himself from his role, while a politician must not do so. He who desires to be successful in politics must inseparably connect succinct personal characteristics, which guarantee a high degree of recognition, with the image of positive deeds (e.g., those of a „committed social activist"). In the public's eyes he then "stands for" those deeds and can utilize this image as a shield in other areas.

But what happens when a politician engages an actor to do his job?

This question is not at all far-fetched, as a number of stories about doubles show.10

Recall the dentist Peter Shapallo, who had the bad luck to be born in the same month as the Albanian dictator Enver Hoxha. As his biographer Lloyd Jones (1993:1ff) tells us, he was "[...] over six feet tall, and broad across the shoulder. The dentist and the dictator had perfect matching shadows. And twin smiles designed to reassure." Height, shoulder width, and smiling lines were the prominent characteristics required for the dentist to be able to play the Great Leader. The rest was taken care of by the plastic surgeon, hairdressers, and tailor employed by the dictator in the 1960's. They modeled the dentist so that the iconic references to the dictator increased, and tried to reduce to zero the number of characteristics having to be indexically inferred. The dictator came into the operation room from time to time and followed the progress of the sign production:

"His glance moved between Shapallo and his own reflection in a hand-held mirror. Once satisfied that the reflection could not be improved, he had Shapallo's family killed -- his wife and two daughters, aged eight and ten. Next to go were the surgeon, hairdressers and tailor. They were in the bus that toppled over the cliffs which spill down to Dhermi on the Adriatic Coast" (Jones 1993:2).

The dictator stuck his double in a cage-like building and let him be taken care of by an attendant who dealt with all of his needs. He was forced to travel behind black-tinted windows and present himself on balconies to the cheering masses at May Day Celebrations; "and once when the leader dreamt of an airplane crash, it fell to the dentist to take the leader's place on a helicopter flight from Vlore to the then Russian naval base nearby" (Jones 1993:4).

Shapallo was the perfect icon of the dictator: "He lost weight when the Great Leader dieted; together their hairlines receded and when the Great Leader sprained an ankle Shapallo limped" Jones 1993:4).

And yet instrumentalizing the double also changed the behavior of the original. To avoid letting any false conclusions arise, the dictator now traveled behind black-tinted windows as well, and his behavior with regard to the population became like that of his double.

"On film, Shapallo is the slow-moving shadow turning to wave to the crowd; there, he pauses from his stride to take a bouquet of flowers from the small girl. Here, he strikes a serious pose. He tilts back his chin and clasps his hands behind his back" (Jones 1993:4-5).

But did the Shapallo in the film think the thoughts of a ruler or those of a dentist? And how often did the dictator regard reality with a dentist's eyes?

It was on the occasion of Shapallo's first public appearance as the Emperor's understudy;

"[...] as his motorcade entered the Tirana Football Stadium, the crowd rose to its feet with a deafening cheer. He was introduced to the players of both teams, and as he moved along their line the players bowed, or smiled so easily or willingly that Shapallo, out of gratitude for their easy acceptance of him in the Emperor's clothes, smiled handsomely back" (Jones 1993:119).

The ruler's sign produced signs to thank the addressees for mistaking it for the original. The addressees took his thanks for a proof of grace of the original.

In certain phases of this drama, the governmental protocol tended to prefer Shapallo's appearances to those of the dictator -- because Shapallo corresponded better to the original's ideal than the original himself. As the ruler lay dying, a sort of cheerful exhaustion was expected of Shapallo. At the insistence of his attendant, Shapallo used a walking stick to support himself.

"On the same advice, these days he allowed himself to be bodily assisted to the rostrum, to take in a parade. He was to smile at the silliness of all the fuss and give every indication that recovery lay just around the corner" (Jones 1993:127).

Both Enver Hoxha and Shapallo were spared the bullet of the assassin. Yet what consequences could such an assassination have had? If the bullet had hit the dictator, then the government might have been tempted to let the dentist simulate convalescence so as to make him take over the full role of the Great Leader. If the bullet had hit Shapallo, then the dictator would have been forced to simulate a wound on his person and a subsequent convalescence. The game would have been risky in either case, because the double was only successful in his work as long as the public knew nothing about him.

It has been established by historians that Enver Hoxha died in the middle of the 80's of Parkinson's disease. The door to Shapallo's cage was deliberately left open on this night so that he could flee. But now the actor was to recognize with a shock that he couldn't convince anyone that he was different from his original. The surgeon, hairdressers, and tailor had done their job too well, and the original himself had taken the last bit of individuality from the image by copying his behavior.

As the burial day dawned, a rumor was going around -- it was said that the Great Leader had been seen:

" [...] like the 'Christ figure' Enver had risen from the dead. The sightings spread out from Tirana to the countryside [...]. Eyewitness reports spoke of a man with 'film star looks'" (Jones 1993:2ff).

How does an iconic sign rid itself of its sign-function when its referent no longer exists? Shapallo tried to reach the German embassy's grounds in Tirana:

"Word passed among the crowds camped along Embassy Row that the ghost of the late dictator had come back to haunt [...] those seeking to leave. There was a terrible commotion. [...] Shapallo was pinned to the fence inside and those on the outside waiting to get into the embassy reached through the fence to rip his clothing. It was left to the Embassy officials to haul the concussed dentist to safety inside the building. A doctor was sent for -- and an ear lobe was sewn back on and several cuts stitched above Shapallo's right eye. [...] A staff member found him the next day, draped over the bathtub, the mirror and bathroom walls splattered with blood and, in the handbasin, the knife which Shapallo had taken to the Emperor's face. Shapallo has lost the tip of his nose. Down the centre of his forehead he's made a deep cut. It was [the German official's] [...] impression that the dentist had tried to peel back the skin" (Jones 1993: 3ff).

The story could only have this shocking conclusion

-- because the distinction between the actor and the person presented, which is always possible in a theater, failed due to a lack of indexical signs;

-- because the audience refused to acknowledge the mistake it had made

in confusing the imitator with the person imitated for years on end;

-- because everyone took the permanent characteristics of a person more

seriously than his shorter utterances.

Due to the predominance of one-sided communication models, the circumstances of humans behaving as signs have not yet been studied sufficiently. Many general problems must be clarified:

-- If people are signs, who can be their sender and who their addressee?

-- If people embody signifiers, who encoded them and who decided about

their proper decoding?

-- What difference does it make whether it is a single person, a type of

person, or a group of persons that becomes a sign for someone?

-- What types of message can be communicated by people who are signs?

-- What makes people capable of having meaning? Is this only achieved in a

society structured into groups of people according to criteria such as sex, age, way of life, occupation, income, background, religion, and race (cf. Posner 1991)?

-- Can the signifieds a person may stand for be explained simply as projections of that person's behavior onto his appearance? Which is more fundamental: the differences in appearance between people or the tendency to semanticize them on the basis of their differences in behavior?

-- Portrayals of people in pictures, tone, and language tend to turn types of people into signs and to suggest to us their meanings. Painting, music, literature, and theater are full of examples. But to which degree do we activate such meaning assignments when we interpret our neighbors in everyday life?

-- Finally, if humans are able to take part in communication not only as senders and recipients but also as signs and signifiers, what is to stop us from analyzing them as channels, codes, or media of communication?

For millenia, man has been known as "homo significans". The time is past in which this was taken to mean only "homo signa faciens" -- sign-making man. In the future, we will again have to take man seriously as "significans" in the narrower sense: as a coded sign -- a sign, certainly, which creates other signs as well.


Angelus Silesius (Johann Scheffler) (1674), Cherubinischer Wandersmann.

In: Hans L. Held (ed.), Sämtliche poetische Werke, 3 vols. Munich: Hanser 1949-1952.

Assman, Aleida (1996), "Probleme der Erfassung von Zeichenkonzeptionen

im Abendland". In: Roland Posner, Klaus Robering and Thomas A. Sebeok (eds.), Semiotics: A Handbook on the Sign-Theoretic Foundations of Nature and Culture. Berlin and New York: de Gruyter: Article 33.

Barth, Karl (1928), "Ludwig Feuerbach". In: K. Barth, Die Theologie und die Kirche (Gesammelte Vorträge, vol. 2). Berlin: Kaiser: 212-239.

Barthes, Roland (1975), Roland Barthes par Roland Barthes. Paris: Seuil.

Cherry, Colin (1957), On Human Communication. Cambridge MA: The

M.I.T. Press. German by P. Müller: Kommunikationsforschung: Eine neue Wissenschaft. Frankfurt/M.: Fischer 1962. 2nd edition 1967.

Deschner, Karlheinz (1974), Das Kreuz mit der Kirche. Eine Sexual-

geschichte des Christentums. Düsseldorf and Vienna: Econ.

Ebeling, Gerhard (1964), "Luthers Reden von Gott". In Schäfer 1964:35-53.

Eco, Umberto (1975), A Theory of Semiotics . Bloominton IN: Indiana

University Press.

Ekman, Paul and Wallace F. Friesen (1975), Unmasking the Face: A Guide to Regognizing Emotions from Facial Clues. Englewood Cliffs NJ: Prentice-Hall.

Feuerbach, Ludwig A. (1841), Das Wesen des Christentums. Leipzig:


Hardy, Donald E. (1992), Forever Yes: Art of the New Tattoo. Honolulu HI: Hardy Marks Publications.

Holy Bible, King James Revised Standard Version. Oxford: Oxford

University Press 1952.

Hucklenbroich, Peter (1996), "Semiotische Aspekte der Medizin:

Medizinsemiotik". In: Roland Posner, Klaus Robering and Thomas A. Sebeok (eds.), Semiotics: A Handbook on the Sign-Theoretic Foundations of Nature and Culture. Berlin and New York: de Gruyter: Article 144.

Ingarden, Roman (1931), Das literarische Kunstwerk . Tübingen: Niemeyer.

2nd edition 1965. English by George G. Grabowicz: The Literary Work of Art . Evanston: Northwestern University Press, 1973.

Jakobson, Roman (1960), "Linguistics and Poetics". In: Thomas A. Sebeok

(ed.), Style in Language. Cambridge MA: The M.I.T. Press 1969:360-377.

Jones, Lloyd (1993), Biografi. London: Penguin.

Kittler, Friedrich (1985), "Romantik--Psychoanalyse--Film: eine Doppel-

gängergeschichte". In: Jochen Hörisch and George C. Tholen (eds.), Eingebildete Texte: Affairen zwischen Psychoanalyse und Literaturwissenschaft. Munich: Fink 118-135. Reprinted in: F. Kittler (1993), Draculas Vermächtnis: Technische Schriften. Leipzig: Reclam: 81-104.

Leach, Edmund R. (1976), Culture and Communication: The Logic by which Symbols are Connected. Cambridge, England: Cambridge University Press.

Li, You Zheng (1996), "China". In: Roland Posner, Klaus Robering and

Thomas A. Sebeok (eds.), Semiotics: A Handbook on the Sign-Theoretic Foundations of Nature and Culture. Berlin and New York: de Gruyter: Article 93.

Luther, Martin (1525), Biblia: Das ist: Die gantze Heilige Schrift (Deudsch).

Auffs new zugericht. Wittenberg: Hans Lufft.

Luther Martin (1529), Großer Katechismus. In: J. Knaake et al. (eds.), Werke (Weimar Edition). Weimar: Böhlau 1910: vol. 30, 1: 123-238.

Meyer-Eppler, Werner (1959), Grundlagen und Anwendungen der

Informationstheorie. Berlin: Springer.

Moser, Manfred (1988), Schönheit, Schnitt und Narbe: Das ästhetische Ideal

der plastischen Chirurgie. Stuttgart: Schwarz.

Peirce, Charles S. (1931-58), Collected Papers. 8 vols., ed. Charles Hartshorne, Paul Weiss and Arthur W. Burks. Cambridge MA: Harvard University Press.

Pollig, Hermann, Susanne Schlichtenmayer and Gertrud Baur-Burkath

(eds.) (1987), Exotische Welten -- europäische Phantasien. Stuttgart and Bad Cannstatt: Edition Cantz.

Ponzio, Augusto (1990), Man as a Sign: Essays on the Philosophy of

Language. Berlin and New York: Mouton de Gruyter.

Posner, Roland (1989), "What is Culture? Toward a Semiotic Explication of

Anthropological Concepts". In: Walter A. Koch (ed.), The Nature of Culture. Bochum: Brockmeyer: 240-295.

Posner, Roland (1995) "Denkmittel als Kommunikationsmittel". Zeitschrift

für Semiotik 17 : 247-256.

Posner, Roland (1991), "Zeichenkultur in Asien" Zeitschrift für Semiotik 13: 3-14.

Prieto, Luis J. (1966), Messages et signaux. Paris: Presses Universitaites de

France. German by G. Wotjak: Nachrichten und Signale. Berlin: Akademie-Verlag 1972.

Probst, Peter (1992), Der dekorierte Körper. Berlin: Museum für Völker-


Rank, Otto (1925), Der Doppelgänger. Eine psychoanalytische Studie. 2nd

edition Leipzig: Internationaler psychoanalytischer Verlag.

Rozik, Eli (1995), "Theatre Notation: Possible and Impossible". In: Sabine

Kowal, Daniel C. O'Connel and Roland Posner (eds.), Zeichen für Zeit: Zur Notation und Transkription von Bewegungsabläufen / Signs for Time: Notation and Transcription of Movements. Special issue of Kodikas / Code . Tübingen: Narr.

Saussure, Ferdinand (1916), Cours de linguistique générale . Lausanne and

Paris: Payot. English by W. Baskin: A Course in General Linguistics. London: Owen 1964.



Göran Rossholm

University of Stockholm






This essay is an attempt to analyze a group of signs, situated in a cross-section between iconicity and indexicality.1 These signs have a common functional feature: they are in some sense used as substitutes for what they stand for. However, as the analysis proceeds, this substitute role will be deemphazised, and their semiotical structure will be put in focus.

Gombrich's Meditations

The art historian E.H. Gombrich's famous essay "Meditations on a Hobby Horse" challenges most ingrained ideas about the representational picture - picture in a sense which includes three- as well two-dimensional objects.2

Gombrich's idea is an evolutionary speculation: representation has deep biological roots and the notion of representation as perceptual resemblance is of later date, and it is misleading, even when it is limited to apply to art from the Renaissance and later. The genetic approach - to tell a Just So Story - separates Gombrich's ideas from most mimetic accounts, also from theories which have been influenced by his essay, such as Kendall Walton's make-believe theory of representation3 - and the present considerations. My point of departure is the kind of symbols which Gombrich focuses on, and how they challenge common ideas of mimesis or iconicity.

Which, then, are the iconic prejudices which Gombrich attacks? The idea of grades of abstraction in representational pictures is questioned, as well as ideas of stylization and of conceptual images. But the central, and undoubtedly the most provocative criticism targets on the foundations of most theories of representation: that pictures are formally similar to what they depict, and that they are referring symbols, that is, a representational picture here and now refers to something else somewhere else, in reality or imagination. A further concept hit upon by Gombrich is the one of communication, the idea that the function of a picture is to communicate something to someone.

Instead of all this Gombrich suggests creation and substitution as mimetic key words. In one of his more slogan-like statements he writes: "We may sum up this Just So Story by saying that substitution may precede portrayal, and creation communication." The second term is the central one: Gombrich says that a child makes a train with some blocks and he puts it in contrast to bills and newspaper photos which strengthen our prejudice that pictures always refer to things elsewhere, in "some imaginary or actual reality". What the child produces is a train, or to put it more cautiously than Gombrich himself does, the utterance "That is a train" about the toy train is more and otherwise justified than the same utterance about a photo of train in an advertisement. The justification comes from the function: the toy train substitutes for a real train, the photo does not.

According to Gombrich the substitutional function is biologically rooted. He enumerates several kinds of objects and actions other than toys and games as examples of substitutional symbols. Egyptian tomb sculptures (horses, servants) serve as the things they represent. A baby sucks its thumb instead of its mother's breast. A cat hunts a ball as if it were a living prey. As can be seen, the examples point backwards, toward the childhood of the icon and even beyond that, in the history of the individual and of the species.

Scheffler's criticism

So far, I have presented Gombrich's ideas without critical comment. The evolutionary speculation - being a speculation - is, of course, open to criticism. However, I am not the right person to comment upon that, and my concern is not with questions about historical, archeological or biological evidence, but with armchair semiotics. And from that point of view Gombrich's use of the relational term "representation", and the consequential opposition between depiction and substitution is certainly questionable.

In an article the philosopher Israel Scheffler has criticised Gombrich's statement that the hobby horse is a horse.4 To put it short: Scheffler claims that if we take this statement literally, it is, of course, absurd; and if we take it metaphorically it is rather trivial. Scheffler's own account runs like this. The hobby horse is, like the word "horse" and like a picture of a horse, a symbol which stands for horses (or for a horse). But we can call the hobby horse a horse mention-selectively, that is, we talk figuratively in such a way that we use the terms of the symbolized instead of the terms of the symbol; we speak about symbols in terms of their denotata.5 Instead of saying "The actor who played Polonius was brilliant" we say "Polonius was brilliant" and instead of saying "The painting of the horse is dilettante" we say "The horse is dilettante". Scheffler traces the figure mention-selection in the child's development of speech. In the first phase the child makes no distinction between symbol and symbolized - the term "horse" applies to horses as well as to horse pictures and horse words. Under pressure from adult semantics the child enters a second phase: the hobby horse is separated from the class of living horses, but only sideways. The hobby horse belongs to another kind of horses, namely pretend horses. In the last, adult, phase horse symbols, including hobby horses, and horses belong to different classes and to different referential strata, but, as a reminiscence of the infantile semiotics we can, as we grow up, still refer to horse symbols with words such as "horse".

Scheffler concentrates his criticism on Gombrich's co-classification of horses and hobby horses, and his conclusion implies another co-classification: hobby horse and the word "horse" and horse portraits belong together.

My own position - which I will develop - is this: the hobby horse as well as the horse portrait is a horse representation, if you like, a horse picture. Still, there are differences between the category of symbols instantiated by the portrait and the category instantiated by the hobby horse, differences in semiotic structure, not only differences in use. These differences will be visible as we observe some trouble which Scheffler's choice of route leads to.


Scheffler's semiotic apparatus is taken over from Nelson Goodman and enrichened by some additions, such as the notion mention-selection.6 I will try to show that a difference in semiotic structure between substitutional and non-substitutional representations is possible to formulate from a Goodmanian point of view.

According to Goodman - and to Scheffler - the conceptual foundation of representaion is denotation: a portrait of Churchill denotes Churchill; the relation is the same as the one between the name "Churchill" and the man. But talk about pictures also includes another aspect: how the picture is classified. A picture of Churchill is, probably, a Churchill-picture. But not necessarilly: a picture of Churchill as a dog is a dog-picture denoting Churchill; denotation and classification part. Pictures representing fictive things denote nothing but are grouped in different classes: Apollo-pictures, Mickey Mouse-pictures. The classification of pictures is, of course, not limited to certain fixed categories. As any other object, a label for a picture can be more or less specified. Our Churchill-picture may be a Churchill sitting in a chair in a garden, bareheaded with a worried expression in his face-picture.

If we apply this terminology to toys, religious and magical representations, dummies, decoys and other substitutional symbols, we get horse-representations, train-representations, servant-representations and, with a generous semiotics, breast-representations and mouse-representations, and so on. However, let us stick to only one example - a doll - and beg the reader to keep in mind the broader repertoire of symbols, hinted at above. The denotational role of the doll can be given three interpretations: it denotes a certain girl, it denotes a girl in general (or the class of girls), or it does not denote at all because it represents a fictive girl. But: in all three cases the doll can be classified as a girl-representation.

The pecularity, which I have announced, has to do with the classification of the doll and with its possible denotata. But first a comment on iconism. Goodman's (and Scheffler's) semiotic theory is iconoclastic; there is no role to play for the concept of similarity. In an article I have tried to formulate a concept of iconism within the Goodmanian theory of symbols, and, hence, without use of relation terms such as similarity or resemblance.7 I will here merely mention one ingredient in this revision of Goodman's theory of representation.

To assert that something, A, represents something, B, in virtue of some features shared by A and B, is trivial, according to Goodman, since every pair of objects share some feature. But, if some of the shared features are incorporated into some structure of reference, linking A to B, the triviality disappears. And this can be done by the use of Goodman's notion indirect (or mediated) reference.8 A round red form may represent a red ball. The colour of the symbol is exemplified, or in a more Goodmanian parlance, the symbol exemplifies a colour label, namely Red. This label, in turn, selectively denotes the represented ball. If the picture is fictional - the represented ball does not exist - the route is a bit more complicated. The picture exemplifies the label Red thing which in turn exemplifies the label Red thing-description, which, finally, selectively denotes the label Red ball - and that is the end of the route. Of course, iconicity cannot be defined in this simple way, but I believe such structures are necessary for representation.

Now to the pecularity. We assume that we have ten dolls in a toy store. All ten look the same to us - the only difference we can discover is their different positions. We classify the ten dolls as symbols with the same referential value - they are synonymous girl-representations.9 We buy one of them and give it to a child. The child puts it on his or her lap and feeds it with a spoon. The doll can now be classified more narrowly: it is a representation of a little girl, who gets food in a particular place - on the child's lap - and at a particular time - let's say now. Irrespective of interpretation - if the doll represents an existing girl, a girl in general, or a fictive girl - we get a classification, which is such that the doll could not denote anything in accordance to its classification. The doll is a Girl-fed-at-place-p-at-time-t-representation, and it could not denote anything because that place at that time is occupied: the doll holds the place of the possible - and, hence, impossible - girl. If we make the thought experiment that such a girl really existed, the symbol would disappear. Sign and signified cannot coexist.


This phenomenon has similarities to verbal deixis: the doll "means" something like Girl Here and Now, and when time goes and the doll moves the references of "Here" and "Now" change. But there is also a difference between words such as "I", "here", "now" and the doll. The deictic word is not annihilated by its denotatum. There is room for both.

To look a little closer at this parallel a few remarks on verbal deixis are needed. As non-deictic descriptive words the deictic ones denote. And like the non-deictic words, the deictic ones follow one and the same principle of substitution: the word or the phrase - in the token sense - can be substituted by any other word - i.e. token - spelled in the same way without any change of reference. In an utterance of mine - still a token - running "I live in Aspudden" the occurrence of "Aspudden" can be replaced by any other occurrence of "Aspudden", and any other occurence of "I" than the actual one can take the subject position without any effect on references or truth-value. But here we can also note a difference between the deictic and non-deictic words. The new token "Aspudden" probably referred to the same thing in its original context, the one from which it was taken. The replacing token "I" on the other hand probably did not denote me in its earlier use.

So far "I" differs from "Aspudden" by being ambiguous. But deictic expressions have a characteristic usually missing in other ambiguous expressions: they refer reflectively, to themselves as tokens. "Here" is the place of the utterance, "I" denotes the speaker of the utterance and so on.

To compare the semantic properties of verbal deixis to non-verbal substitutional symbols the type-token distinction must be extended to cover objects such as toys. There are certainly considerable difficulties in doing so, but to make my point I will, without discussion, simply assume that the ten dolls are tokens of the same type. The first question is, then, whether the doll held by the child is substitutable by a doll from the store without change of meaning. If we classify the first doll as a Girl-at-place p-at-time t-representation the new doll can probably take the place of the old one without change of classification. On the other hand, if the label runs The girl Anne-at p-at t-representation, then it is quite reasonable to expect that the new doll cannot be so classified; she is probably not Anne. The rest of the deictic properties match the doll better. If a doll from the store takes the original doll's place, the former changes its semantic value to obtain one identical to the original doll's (such as being a Girl-at p-at t-representation), a shift similar to what happens to the cut-out "I" in its new environment. And our doll has not the same semantic value as other tokens of the same type: one doll may be a Sleeping girl-at p-at t-representation, another may be a Crying girl-at p-at t-representation, and so on. And finally, the doll refers to something related to itself, the place and time that it occupies; as time goes and the symbol moves the references change.

A remaining difference is the pecularity I have mentioned repeatedly: if the doll is assumed to denote deictically, as the phrase Girl here and now, the denotatum takes the place of the symbol, and consequently is no denotatum of that symbol.

Index and indirectness

So far we have related the doll to only a special kind indexical symbol: verbal deixis. However, there are others; the lectures presented at this conference bear witness to that. The linguist John Lyons separates deixis from index in a way which fits our doll.10 His example is an American speaking English with an American accent, thereby indicating that he is an American. In Goodman's terms you could say that he exemplifies the label American. But there is a minor difference between Lyons' example and Goodman's most frequent specimen of exemplification: the tailor's swatch. A bit of textile exemplifies some properties, such as colour and texture, but the choice of exemplifying symbol lacks significance. Any bit of textile with the requested properties will do. The American is not substitutable in the same way: he is not only indicaticating that he is an American, he also indicates that he is an American; he is expressing his identity. The symbol is the man (and his speech), and the exemplified label may be phrased Is identical to the symbol and is an American. Both verbal deixis and index involve self-reference but in different ways. The token "here" in an utterance denotes a place related to the symbol itself, and thereby denotes itself. Index, in Lyons' sense, refers in the opposite direction, to something which denotes the symbol - not to something which is denoted by the symbol - and which in turn refers - by denotation - to the symbol.

Of course the parallel to the doll is not strict: the doll does not express its identity - the doll is not a girl - but only its pseudoidentity. Literally speaking, the doll exemplifies a term which iconically attributes girlishness to the doll, a term which in turn via another mediating term of a higher order, which in turn selectively denotes a label such as Girl at place p at time t. In other words, the referential structure is the one presented above as the indirect reference pattern of fictional icons.

Let us return to question left answered: is Scheffler correct in equating the figurative status of three different statements running "A is a girl" about 1) the word "girl", 2) a photo of a girl and 3) a doll? A graphic summation shows that there is more to 2) and 3) than mention-selection, and that there is a difference between 2) and 3):

(Arrows running upwards indicate exemplification, downwards denotation. The letter "A" abbreviates those iconic properties which make us see the picture and the doll as girl-representations.)

The metaphorical movement in the first case is only vertical: we move the term "Girl" one step down and call both the human being and its description "girl". In the second and the third case the movement is sideways: we exchange the literal description of the symbol - "A" - by the final label - "Girl" - but these two labels are on the same level, and they are, in that sense, closer to each other.11 And the literal and figurative label in the third case are closer to each other than in the second case: the references to the time and the place of the symbol are preserved through the route. Of course Scheffler is correct in his criticism of the literal interpretation of Gombrich's statements that the hobby horse is a horse and that the toy train is a train. But it seems justified to distinguish a sense in the figurative interpretation, which adds something to Scheffler's mention-selection, and which also differs from the non substitutional cases of iconic representation. And it also seems to me reasonable to speak of differences in strength between the three cases; the figurative statement that the child holds a girl, not a doll, is closer to a literal truth - at a certain place, at a certain time something is, in some restricted sense, acted upon and taken as a living girl.

Final remarks

I have used the term "substitutional symbol" to refer to the kind of objects Gombrich has called attention to in his "Meditations", and this phrase is in accordance with Gombrich's ideas. However, in my account all references to substitution have disappeared, and the things studied are not certain objects but objects used in a certain way, and, finally, this use cannot be defined as substitution without further elaborations.

Let us, for the last time, return to the doll, and let us assume that the child not only plays a game, where the doll acquires a pseudoidentity, but where the surroundings - the time and the place - and the child itself are given new make-believe properties. The child's pretense is this: he or she, who is no longer him/herself but someone else, feeds a little girl and they are both in a little cottage in the middle of a forest a long, long time ago. This is, as much as the earlier account, an example of playing with a doll, but it is not an activity, where the constituents can be analysed as in 3) above; the references to here and now have dissappeared. In spite of the functional similarities, this is a kind of symbolism closer to the picture (or a theatre performance), as analysed in 2) above. Thus, the dividing line between things à la 2) and things à la 3) is whether or not the symbol should be read as embedded in its real environment (or as embedded in what is believed to be its real environment).




Göran Sonesson

University of Lund






When terms like "index" and "indexicality" are invoked, authors tend to take one out of two, equally strange, points of view: either they simply ignore the fact that, historically, Peirce is responsible for introducing at least the former term (and perhaps some corresponding concept) into our vernacular; or, like some old Talmud scholar reading the Scriptures, or a Marxist adept of the sixties, they are exclusively intent upon finding out what Peirce "really meant", without making any distinction between the internal philological truth, and that which is true in the world.

Here, I will not be content with finding out "what Peirce really said", nor will I just extract out of the entire body of the text that which I happen to need, but I will try to reconstruct Peirce’s conception with a view to fairness, and then go on to criticise it, from the standpoint of what appears relevant to present-day semiotics. In so doing, however, I will suggest a distinction between "index" and "indexicality" which is not directly justified in Peirce’s writing, though it may be derived from the latter. This will have the effect of making indexicality into a property of perception generally.


The index in its triadic home

In Peirce’s work, the index appears, together with the icon and the symbol, as a member of one of the numerous triads profusing the world of our experience. In this particular case, the division is made on account of the particular character of the relation joining expression and content ("representamen" and "object" in Peircean parlance). As far as I have been able to tell, Peirce’s formulations are compatible with two quite different conceptions: on one hand, there may be three properties, iconicity, indexicality, and symbolicity, which, by themselves and without any further requirement, trigger off the recognition of something as a sign, and at the same time as a particular kind of sign. On the other hand, something that is already, for some other reason, recognized as being a sign, could be discovered to be an indexical sign, rather than an iconic or symbolic one, by means of tracing it back to the indexical ground. In the case of iconic signs, it should be clear that the counter-arguments which have been formulated can only apply to the former case; yet we have argued elsewhere that, in least in the case of iconicity, both conceptions are valid and serve to distinguish different types of iconicity (Sonesson 1989a; 1996ab, to be published).

In the following, I will take the latter conceptions for granted: thus, indexicality can be conceived as a property which makes something which is a sign into an index. However, by a slight shift of emphasis, it could be construed as a property which, when added to the sign function, creates an index, but which, in addition, may have other parts to play in the constitution of meaning. Such a conception might account for the ambiguities of the Peircean notion, as well as for some of the uses to which it has been put subsequently.

Given the long period through which Peirce’s thinking evolved, and the state in which it came down to the public, it is not surprising that indexicality, like so many Peircean notions, should be so variously, and inconsistently, defined, and that many of the examples given hardly fit in with the definitions (cf. Goudge 1965; Sonesson 1989a; 1996abcd).

In this context, it is important to show a certain "fairness" to Peirce: however contradictory his various definitions may appear, in between themselves, and in relation to the examples given, we should at least try to make sense of the implicit structural argument contained in Peirce’s theory: indexicality must be understood in such a way that, together with two other relationship, viz. iconicity and symbolicity, it will exhaust the domain of signs. In order to make sense, the division must depend on the same criteria in all cases, viz. on the nature of the relationship between expression and content, and more precisely, that which serves as a motive for our joining them together (contrary to what is the case with the extended lists proposed by Sebeok and Helbo). Thus we should interpret indexicality, as well as iconicity and symbolicity, in such as way that no fourth sign type appears to be needed, at least until the whole enterprise has been shown to be unfeasible.


Iconic and indexical grounds

For those who are no true believers, it may be difficult to make sense of Firstness, Secondness, and Thirdness. And yet it will be necessary to start out from these logical rudiments of Peircean metaphysics. Indexicality pertains to the general category of Secondness, which means that it concerns two items and/or the relation between them. The sign being a Third, there is every reason to think that it cannot be constituted by indexicality alone. Perhaps Peirce is really considering "potential sign-vehicles" in order to investigate their "capacity to serve as signs" (Bruss 1978:87). More substantial arguments for this may be derived from a consideration of the Peircean concept of "ground", which is useful for understanding the nature of indexicality, although it seems to disappear in later texts

In one of his well-known definitions of the sign, or rather the sign-vehicle, Peirce (2:228) describes it as something which "stands for that object not in all respects, but in reference to a sort of idea, which I sometimes called the ground of the representation". According to one of his commentators, Greenlee (1975:64), the ground is that aspect of the referent which is referred to by the expression, for instance, the direction of the wind, which is the only property of the referential object "the wind" of which the weathercock informs us. On the other hand, Savan (1976:10) considers the ground to consist of the features picked out from the thing serving as expression, which, to extend Greenlee’s example, would include those properties of the weathercock permitting it to react to the wind, not, for instance, its having the characteristic shape of a cock made out of iron and being placed on a church steeple. In one passage, however, Peirce himself identifies "ground" with "abstraction" exemplifying it with the blackness of two black things (1.293). That, of course, would be an iconical ground (such as the similarity of the weathercock to a real cock); an indexical ground, in a parallel fashion, would then be whatever it is that connects the properties of the weathercock as a physical thing to the direction in which the wind is blowing. The ground, it appears, is a part of the sign having the function to pick out the relevant elements of expression and content. Therefore, the ground is really a principle of relevance, or, as a Saussurean would say, the "form", connecting expression and content (cf. Sonesson 1989a:205ff).

Generally put, an indexical ground, or indexicality, would then involve two "things" that are apt to enter, in the capacity of being its expression and content (i.e. "representamen" and "object"), into a semiotic relation forming an indexical sign, due to a set of properties which are intrinsic to the relationship between them, such as it is independently of the sign relation. This kind of ground, which is a relation, is best conceived in opposition to an iconic ground, which really consists of two sets of properties which happen to be of the same kind, and the symbolic ground, which is a non-entity, since the motivation of the sign has no existence independently of the sign itself. This is the sense in which indexicality is Secondness, iconicity Firstness, and symbolicity Thirdness.

To be more precise, it would appear that, in Peirce’s view, two items share an iconic ground, being thus apt to enter, in the capacity of being its expression and content, into a semiotic function forming an iconic sign, to the extent that there are some or other set of properties which these items possess independently of each other, which are identical or similar when considered from a particular point of view, or which may be perceived or, more broadly, experienced as being identical or similar, where similarity is taken to be an identity perceived on the background of fundamental difference (cf. Sonesson 1989a,III.1-3.).

Contrary to the indexical ground, which is a relation, the iconic ground thus consists of a set of two classes of properties ascribed to two different "things", which are taken to possess the properties in question independently, not only of the sign relation, but of each other. Indexicality as such involves two "things", and may therefore be conceived independently of the sign function. Since iconicity is Firstness, however, it only concerns one "thing". Indeed, as Peirce (3.1.; 3.362; 4.447) never tires of repeating, a pure icon cannot even exist: it is a disembodied quality which we may experience for a floating instant when contemplating a painting out of awareness. Perhaps, then, to use some of Peirce’s own examples, the blackness of a blackbird, or the fact of Franklin being American, can be considered iconicities; when we compare two black things or Franklin and Rumford from the point of view of their being Americans, we establish a iconic ground; but only when one of the black things is taken to stand for the other, or when Rumford is made to represent Franklin, do they become iconic signs (or hypo-icons, as Peirce sometimes said). Just as indexicality is conceivable, but is not a sign, until it enters the sign relation, iconicity has some kind of being, but does not exist, until a comparison takes place. In this sense, if indexicality is a potential sign, iconicity is only a potential ground.

In sum, then, iconicity begins with the single object; indexicality starts out as a relation. The problem, therefore, consists in determining what kind of relation it is.


Contiguity as opposed to causality

Such a view of indexicality as the one reconstructed above best fits in with the most general formulations given by Peirce, according to which it depends on there being a "real connection", an "existential relation", a "dynamical (including spatial) connection" and even, in one of its many conceivable senses, a "physical connection" between the items involved (Peirce 1.558; 1.196; 2:305; 3.361; 8.335). From this point, it seem natural to go on to argue that indexicality is involved with "spatiotemporal location" (Burks 1949:683ff), which underlies the "indices" of such logicians as Bar-Hillel and Montague, the "egocentric particulars" of Russell and the "shifters" of Jespersen and Jakobson. In fact, however, as Savan (1976:25ff) observes, location in time and space will only result, to the extent that some system of co-ordinates has been conveyed by other types of signs — or, as I would add, to the extent that it can be presupposed by the ongoing practice of the ordinary world of our experience, the world taken for granted, our common Lifeworld.

More generally, many of the examples adduced by Peirce would justify us in going along with Jakobson (1979), when he claims that indexicality is based on "real contiguity", and is connected with the syntagmatic axis of language, and the rhetorical figures of metonymy. To Jakobson, however, metonymy actually involves, not only the relation of contiguity of traditional rhetoric, but also that of part to whole, known in rhetoric as synecdoche. This distinction may be reestablished inside the category of indexicality (cf. Nöth 1975:20f), and could be described more generally in terms of contiguity and factorality (cf. Sonesson 1989a:40ff).

There is, however, another series of definitions which suggest that indexicality is, is some way, dependant on there being a relation of causality between the expression and content of the potential sign: that is, the index supposedly "denotes by virtue of being really affected by that object" (2.248). Apart from this, Peirce also makes a number of other claims about indices, many of which are repeated by Dubois (1983: 48f, 60ff) when trying to demonstrate that photographs are indices: that they refer to unique, singular objects (2.283); that they testify to the existence of its object (2.316); that they show up the object without asserting anything about it (3.361); and that they point, by "blind compulsion", to the object of reference (2.306).

Although the definition by causality is probably the most commonly quoted of all the definitions Peirce offers of indexicality, it has come in for serious criticism. Some commentators would reject the relation between causality and indexicality altogether, while others would see it as merely coincidental. Burks (1949:649ff) takes Peirce to task for confusing the semiotic relation with mere causality, when treating, for instance, the weathercock, which is causally affected by the wind, as an instance of indexical signs: it is not clear, however, why causality should preclude indexicality, since the fact of the wind causing the weathercock to turn must be seen be the observer to be a contiguity in order for it to receive an interpretation.

More to the point, Goudge (1965:55) claims that not all examples of indexical signs given be Peirce are susceptible of receiving a causal explanation: the Pole Star, for instance, may be an index of the north celestial pole, but it is in no way caused by that astronomical location. Nor is a personal pronoun, or even a pointing finger, as I have argued elsewhere (Sonesson 1989a:39, actually caused by the person or thing for which it stands; and if they may be said to motivate it, then this is also true of all other signs. Moreover, if could be added that even some cases which are often taken to confirm the causal explanation are actually doubtful: the causal agent may not be that which is signified, or may not signify in the same respect in which it is the cause. Of all the innumerable causes that have to concur in order for a rap on the door to occur at a particular moment, the door and the material of which it is made, and a particular person and his moving hand may seem to be the most important. However, if, at this moment, no person in particular is expected, the sign will only carry some very general meaning such as "there is somebody (probably a human being) outside the door who wants me to open it and let him in". Neither the particular person, nor his hand or the door, which are the causal agencies, are here parts of the meaning of the sign (Sonesson 1989a:39).

The idea that indices must point to their object by "blind compulsion" could be taken as a special case of causality, this time applied to the interpreter, and thus more properly described as motivation. Greenlee (1973:86) believes this to constitute a contradiction on the part of Peirce, since the interpreting mind is on the level of Thirdness, and thus lies outside the definition of indices, which derives from Secondness. It seems, however, that the contradiction, if there is one, should be located at another point, for already the "immediate object" must (perhaps contrary to the "dynamical object") be a mental unit. There is certainly an extremely Pickwickian sense in which all indices force us to attend to their objects, but in that sense the observation applies to all signs, and even to other kinds of meaning).


Indices and indicators

This brings us to another, rather common, confusion: that between indices and indicators. The term chosen by Peirce certainly suggests that all indices, like the pointing index finger, or an arrow, serve to pinpoint a particular object, to isolate it and bring it out of the, typically spatial, context into which it is ordinarily enmeshed; and this is indeed what Peirce affirms (3.361; 4.56).

However, if we use the term indicator to describe signs which are employed to single out an object or a portion of space for our particular attention, it may be argued that they are not necessarily indices in Peirce’s sense, and that they are not, in any event, sufficiently characterised by being so classified (cf. Sonesson 1989b:50ff, 60f, Goudge 1965: 65ff). Thus, certain indicators, such as pointing fingers and arrows, do suppose a relation of contiguity with that which they point to; but this is not necessary, or even possible, in the case of many verbal indicators, most maps, and the photographic options depending on film, lighting, and frame described as indexical in the semiotics of photography; for, in these cases, the indicative gesture is merely recreated at the level of content. At least some of these examples would also be described by Peirce as not being "genuine" indices.

On the other hand, real indicators, such as fingers and arrows, are equally contiguous to a number of objects which they do not indicate, for instance to the things which are at the opposite side of the arrow-head, in the direction to which it does not point (Sonesson 1989a:47). Therefore, something beyond mere indexicality is required, in the case of the arrow, for instance, the forward thrust of the arrow-head as imagined in water, or the sentiment of its slipping from our hands, as Thom (1973) has suggested. To term certain signs "indicators" is, obviously, to make a categorisation of signs on the basis of their functions, as seen in relationship to the over-all scenes in which signs are produced. We should not expect this categorisation to coincide with the one stemming from Peirce’s classification, which depends on the nature of the relationship between the expression and the referent or content of the sign. Of course, from this point of view, the term "index" is a misnomer, for although the finger so termed may function as an index, its specific function goes beyond that.

Another sense of index current in semiotical literature, which has no obvious relation to that of Peirce, is the one employed by Prieto (1966:15ff; in French "indication"). In order to describe how signs come into being, he imagines an elementary situation in which a set of hoof-prints made by a horse may be observed. Something which Prieto calls the "significative indication" tells us that there is a horse around. There is not, in this case, as there would be in a linguistic sign, any "notificative indication", which would convey to us the idea corresponding to the phrase "attention! this is intended to contain a message". In this particular case, the hoof-prints would be indices also in Peirce’s sense, but not for the same reason: not because they are non-intentional, but because of being markings on a surface. In the case of a linguistic sign, however, the significative indication, which corresponds to the perception of the sound, would not ordinarily be indexical.


The singularity of perception

Indicators normally concern singular objects, as I noted above. Indeed, it is also part of the claim made by Peirce (2.306) that all indices refer to a singular instance, not to some general category. Objections to this generalisation easily present themselves. From the size of an imprint left on the ground it may be possible for the interpreter to determine that the animal which has passed by is a horse, rather than a donkey, but normally there would be nothing in the expression of this index itself permitting him to determine the identity of the horse in question, although, if he knows that there is only one horse and one donkey inside the fence, he can draw a plausible conclusion as to which individual animal is involved. It might be argued, of course, that in any case, only one, particular, animal left the imprint; but the case is quite similar to the knock at the door, where, although a particular person must be doing the knocking, the knock itself merely means "there is someone on the door", unless we possess some additional information. The same argument may be applied to the photograph, in particular to the photogram, in which the referent would not normally be recognisable (cf. Sonesson 1989b:59ff).

Goudge (1965: 60f) also argues against this generalisation, quoting the case in which a demonstrative pronoun ("that") refers to Newton’s First Law, which as such, is not a singularity. Outside the linguistic domain, other interesting examples can be found. According to Peirce, the rolling gait of a man is an index of his being a sailor: but being a sailor is a social role, not a singularity. More importantly, however, the gait is part of a social habitus defining this role, which makes it into a part of a whole (a factorality). But if the relationship of a property to that of which it is a part is indexical, then it is reasonable to think that indexicality will also account for the relation between an item and the class of which it is a member. Such examples are apparently not among those mentioned by Peirce, but they are often cited by later semioticians: thus, for instance, if a pretzel is an index of a bakery (cf. Norrick 1983:230f), than that must be in virtue of its being a member of the class of products sold in the bakery. A class is of course not a singular object, but it may be considered a collection of objects. Often, however, such a class is itself determined by abstract properties. A tailor’s swatch, for instance, is a sign of a class of cloth having the same quality and pattern, but not the same shape or size. Some samples, for instance colour samples, may even be indices of abstract properties themselves (Sonesson 1989a:43ff, 137ff; 1989b: 60f).

Here the importance of what we called the structural argument comes to the fore: if we are going exhaust the domain of signs admitting only three kinds of relationships, neither causality nor singularity can be defining characters of indexicality. This also applies, as we shall see, to several other properties usually ascribed to indices.

In order to consider whether indices demonstrates the existence of their object, it might be necessary first to discuss the meaning of existence (cf. Goudge 1965:58ff). However, if existence is taken to imply the physical occurrence in the ordinary world of our experience, it does not seem to apply to all indices, not, for instance, to the cases considered above, in connection with singularity. A person having the rolling gait of a sailor may, in fact, not be a sailor; and the pretzel hung out above the bakery (admittedly an icon of an index) is still to be seen when the bakery is closed, and no bakery products are for sale. Plausible indices of a unicorn may be produced using a set of horseshoes and a bull’s horn, and do not testify to the existence of unicorns. A faked photograph of a unicorn, or whatever, may be assembled, using pieces of real photographs, processing them in a computer, or even creating them entirely by means of a computer program. Of course, the latter pictures are no photographs, and so no indices, but there is no way we can discover that from looking at them (cf. Sonesson 1989b:61f). For all practical purposes, then, indices cannot testify to the existence of their objects.

As discussed above, indexicality emerges as a potential sign, or, better, as a particular kind of ground characterising indexical signs, but which may also be found outside signs. Perception would seem to be profused with indexicality. Indeed, proximity is a basic factor of perception according to Gestalt psychology, and is also one of the relationships included in topological space perception. The relation of part to whole is fundamental to Gestalt relations themselves. All indexical relations involve either contiguity or factorality. Those indexicalities which are not as yet signs, being based on items which are not situated on different levels of directness or thematisation, or not clearly differentiated, may be described as contexts (or ‘pairings’, in Husserl’s sense). Any experience of two elements being related by proximity, conceived as a primordial perceptual fact, may be considered an actual perceptual context involving contiguity. A actual perceptual context involving factorality is any experience of something as being a part of a whole, or as being a whole having parts (cf. Sonesson 1989a,I.2.5).

When only one of the items is directly given, and the other precedes it in time, or follows it, we may speak of an abductive context (protention and retention, respectively). The term abduction is employed here in Peirce's sense, to signify a general rule or regularity which is taken for granted and which links one singular fact with another. As opposed to deduction and induction, abduction (or "hypothesis" as Peirce first called it) "is where we find some very curious circumstance, which would be explained by the supposition that is was a case of a certain general rule, and therefore adopt that supposition" (2.624). All experience taking place in time is of this kind, for instance our expectancy, when seeing the wood-cutter with the axe raised over his head, that on the following moment, he is going to hit the piece of wood (contiguity protention), and that on the moment just preceding he lifted the axe to its present position (contiguity retention). Indeed, Peirce’s principle is known quite independently in perceptual psychology as the theory that perception consists in hypothesis-testing (cf. Sonesson 1989a, 30ff, 251ff).

Abductive contexts involving factorality would be, using some Peircean examples, the gait of the sailor, the symptom as part of the disease, part and whole in a picture, the partly destroyed Minoan fresco, a jig-saw puzzle, a piece of torn paper (the last three examples combine factorality and contiguity). We may use the term proto-index for an indexicality which is only momentarily a sign, as would be the "tableau vivant" of the wood-cutter, the photographic pose (which is a limitation in time), that what is seen in the view-finder (with spatial limits), and indeed many of the examples given above, to the extent that the flow of indexicalities is momentarily halted.

All perception is profused with indexicalities. From this point of view, the archaeologist’s craft consists in transforming indexicalities of decayed cultures into proto-indices and indices accessible to us. The signs which he reads out aloud for us were merely indexical grounds in the ongoing practice of a bygone world.


From indexicality to index

This is a propitious moment to recall the very different use of the term "index" made by Piaget (1967) when introducing his idea of the semiotic function (which, in the early writings, was less adequately termed the symbolic function), which is a capacity acquired by the child at around 18 to 24 months of age, enabling him to imitate something outside the direct presence of the model, to use language, make drawings, play "symbolically", and have access to mental imagery and memory. The common factor underlying all these phenomena, according to Piaget, is the ability to represent reality by means of a signifier which is distinct, or differentiated, from the signified. Even before that age, however, Piaget believes that the child is able to "connect significations" by means of "indices" and "signals", which do not suppose any such differentiation between expression and content. The signifier of the index is, Piaget says, "an objective aspect of the signified"; thus, for instance, the visible butt of an almost entirely hidden object is the signifier of the object for the baby; and the tracks in the snow stand for the prey to the hunter, just as any effect stands for its cause. But when the child uses a pebble to signify candy, he is well aware of the difference between the two, which implies, as Piaget tells us, "a differentiation, from the subject’s own point of view, between the signifier and the signified".

It is important to note that, while the signifier of the index is said to be an objective aspect of the signifier, we are told that in the sign, expression and content are differentiated form the point of view of the subject. We could actually imagine this same child that in Piaget’s example uses a pebble to stand for a piece of candy having recourse instead to a feather in order to represent a bird, without therefore confusing the feather and the bird: then the child would be using the feature, which is objectively a part of the bird, while differentiating the former from the latter from his point of view. Only then would he be using an index, not in the sense of Piaget, but in that of Peirce and most semioticians, rather than a mere indexicality; and obviously the hunter, who uses the tracks to identify the animal, and to find out which direction is has followed, and who does this in order to catch the animal, does not, in his construal of the sign, confuse the tracks with the animal itself, in which case he would be satisfied with the former (Sonesson 1989a:50f).

As the term is employed by Piaget, the index is thus not a sign: it may therefore better be termed an indexicality, or an indexical ground, which as yet has not attained the status of sign. When suitably defined, however, Piaget’s notion of differentiation may be used, together with other criteria, to distinguish signs from other meanings, and thus to tell indices and other indexicalities apart. The other criteria could be adopted from Husserlean phenomenology. In fact, while both Saussure and Peirce simply take the notions of expression and content (no matter what they call them) for granted, Piaget sees that their distinction must emerge in time: and only Husserl has described their difference in terms of directness and focality. Following Husserl, the expression may be said to be directly given but not in focus, whereas the content, which is the focal member, is given only indirectly (cf. Sonesson 1989a:50f).

The lack of definitions may explain that Peirce tended to over-extend the notion of sign. In his later days, however, he realised that all his notion were to narrow: instead of sign, he should have talked of mediation, and the latter should be understood as branching, that is, as a crutch (Cf. Parmentier 1994). We could perhaps take the third part of this crutch to be the point of view of the subject, of which Piaget talks (the place prepared for him in the sign is of course the famous "interpretant"), and although the two others parts cannot be said to be clearly differentiated, the perspective of the subject certainly spans them both. In this sense, the ground is already mediation, although it is not a sign; and it starts out as perceptual mediation.


On several kinds of indices

Apart from being a sign, an index, in the Peircean sense, must, as we have seen, contain an indexical ground . The fact that such a ground could exist independently of the sign relation should not be taken to mean that the indexical relation necessarily has to precede the sign relation in real time. Indeed, some indexical relations must come into being at the same time as the sign is produced, as is the case, for instance, of verbal "shifters": the person indexically related to the sign "I" is the one which at a particular moment pronounces the sound /ai/, which is to say that the indexical ground is produced at the very same moment in which the sign is put to use. Similarly, there is no class of "pointed-out objects" known to exists, but a member of such a class is created each time an act of pointing takes place (see Sonesson 1989a,I.2.5.).

Many of Peirce’s own examples of indexical signs are of the kind which acquire their meaning thanks to a regularity which is known to obtain between different facts. Since a kind of reasoning which connects two facts by means of supposed regularity is called an abduction by Peirce, we might perhaps call this group of signs abductive indices. They can involve contiguity, as in the case of footprints, fingerprints, the cross as a sign of the crucified, the weather-cock (contiguity to the direction of the wind); or factorality, when an anchor is used to stand for navigation, the clock to designate the watch-maker's (as part of the sum total of clocks), or a painting to indicate the painter’s workshop.

Some of Peirce’s examples, and many of those suggested later, are however of another kind, for, instead of presupposing a regularity known to obtain between the "thing" which serves as expression of the sign, and another "thing" which is taken to be its content, they transform something which is contiguous, or in a relation of factorality, to the expression, into its content. These signs may therefore be termed performative indices. With contiguity, they give rise to such phenomena as the pronoun "you", the finger pointing to an object, the weathercock (as marking the here-and-now of the wind), the clock of the watch-maker's (as marking the emplacement of the shop); and with factorality, they may produce the pronouns "I", "here", "now", the finger pointing out a direction, etc.

Finally, many secondary signs, or signs standing for other signs, are often said to be indexical. Secondary indexical signs are signs, the dominant sign relation of which is pictorial or otherwise iconical, conventional, or whatever, but which involve a secondary sign function, which is a relation between several signs, or parts of several signs, which is indexical. The most obvious case of such signs are the rhetorical figures of metonymy and synecdoche, in which the entire primary sign is related to another sign by means of the respective contents. More trivial are such examples in which the primary sign content is related to a secondary content outside the sign (implication, "connotation" in Eco's sense, metonymy and synecdoche in a loose sense, etc.), some examples of which are dead rhetorical figures such as the sword for the army, a picture of a cross for the crucified, the traditional "symbols" of iconography (all with contiguity); dead rhetorical figures such as the sail for the ship, a picture of a clock for the watch-maker's, iconography generally (all with factorality).

Another variety, which is often parasitic on pictorial signs, appears when the entire primary sign, which is related to another sign via their respective expressions, form an actual perceptual context. This type is often found in publicity and in surrealist painting (cf. Nöth 1975; 1977; Williamson 1978). Some examples, with contiguity, would be two figures seen against a common ground, for instance an advertisement for a brand of whisky with a glass and a bottle; a bottle of gin and a crown; a jetty and a tyre; or pictures placed side by side as in collages, montages, etc. With factorality, examples would be a pile of fruits forming a crown; Magritte’s drawing of a face which is also a female trunk; slices of orange forming a jam bottle; Arcimboldo’s portraits, etc. Other varieties may also be distinguished (cf. Sonesson 1989a,49ff).

Peirce himself introduces a distinction between different kinds of indexical signs, but in terms which are not easy to interpret: thus, he claims that there are designations, which stand for "things or quasi-things with which the interpreting mind is already acquainted", and reagents, which serve "to ascertain certain facts", based on our knowledge of the "connection with the phenomena it indicates" (8.368n). Examples of the first kind are personal, demonstrative and relative pronouns, as well as proper names; an instance of the second kind, however, is when "water placed in a vessel with a shaving of camphor thrown upon it will show whether the vessel is clean or not.". Goudge (1965:55f) and Greenlee (1973:86f) interpret this as a distinction between genuine, causal indices and degenerates cases, but there could also be a suggestion of the distinction, formulated above, between abductive and performative indices. If so, however, Peirce’s wordings are somewhat unfortunate: for while we may have to be acquainted beforehand with the person to which we apply the term "you" a perceptual object, we do not need to know anything about his "youness": it is produced in the act of talking about him (Sonesson 1989a:54).

We do have to know beforehand about the relations to be expected between water, camphor, and uncleanness, just as we have to know about the relation between the clock and the watchmaker’s. But the case considered by Peirce also contains information that this particular vessel is unclean. Thus, there is causality, as well as singularity, and a message is produced at the very same moment at which the sign is produced. However, there could be a class of "unclean vessels", unlike a class of "yous", independently of the sign referring to them. Both classes appear to be limited in time, but only the second is limited by the workings of the sign function itself. Therefore, we may conclude, at least, that the example given by Peirce is too complex to offer any useful division of indices into sub-types.



In this essay, I have tried to liberate the notion of ground from its absorption into the notion of sign, and put it to other uses permitting us to understand the more elementary practice of interpretation in the our common-sense Lifeworld. In preparation for my argument, I have observed that Peirce’s famous trichonomy of icons, symbols and indices may be understood either as listing three properties which singly define something as a sign, or as properties ascribed to something which is for other reasons known to be a sign. I have also had recourse to what I have called a structural argument, according to which the definition of, for instance, the index, must be sufficiently broad for it to exhaust, together with the two other properties, the domain of all conceivable signs.

On these premises, I have demonstrated that indexicality cannot by defined by causality, but must depend on some more general property of real connection. I have also voiced doubts on some other properties commonly associated with indices, such as singularity and testimonial power. The resulting notion of indexicality was then shown to be so general in import as to underlie all our doings in the ordinary Lifeworld, which is first and foremost a world of perception. Finally, I have argued that the independence of the indexical ground from the sign function does not imply that the former must precede the latter in time. Indeed, two kinds of indices may be distinguished on the basis of the temporal relation to the sign function: the abductive index, which presupposes an earlier connection, and the performative index, which causes the connection to which it refers to obtain in the universe of our interpretations.




Bruss, Elisabeth, "Peirce and Jakobson on the nature of the sign", in The sign — Semiotics around the world. Bailey, R.W., Matejka, L., & Steiner, P., (eds.), Michigan Slavic Contributions: Ann Arbor; pp. 81-98.

Burks, Arthur, "Icon, index, symbol", in Philosophy and phenomenological research, IX:4, 1949, pp. 673-689.

Dubois, Philippe, L’acte photographique. Paris & Bruxelles: Nathan/Labor 1983.

Goudge, Thomas A. "Peirce’s Index", in Transactions of the Charles S. Peirce Society, Amherst, Mass.; 1965, 1, pp. 52-70.

Greenlee, Douglas, Peirce’s Concept of Sign. The Hague & Paris: Mouton, 1973.

Husserl, Edmund, Zur Phänomenologie des inneren Zeitbewusstseins Halle 1928: The Hague: Nijhoff 1966.

Jakobson, Roman, "Coup d’œil sur le développment de la sémiotique", in A semiotic landscape/Panorama sémiotique. Actes du premier congrés de l’Association internationale de sémiotique, Milan, juin 1974. Chatman, S, Eco, U., & Klinkenberg, J.M., (eds.); The Hague, Paris & New York Mouton 1979; pp 3-18.

Norrick, Neal, "Principles of motivation", in Semiotics unfolding. Proceedings of the Second congress of the International Association for Semiotic Studies, Vienna, July 1979, Borbé, Tasso, ed., Berlin, New York & Amsterdam: Mouton 1983; volume I, pp. 229-236.

Nöth, Winfried, Semiotik. Tübingen: Niemeyer Verlag 1975.

Nöth, Winfried, Dynamik semiotischer Systeme. Stuttgart: Metzlersche Verlagsbuchhandling 1977.

Parmentier, Richard J., Signs in Society. Bloomington: Indiana University Press 1994.

Peirce, Charles Sanders, Collected Papers I-VIII. Hartshorne, C, Weiss, P, & Burks, A, (eds.). Cambridge, Mass. Harvard University Press 1931-58.

Piaget, Jean, La psychologie de l’intelligence. Paris: Gallimard 1967.

Prieto, Luis, Messages et signaux. Paris: PUF 1966.

Savan, David, An introduction to C.S. Peirce’s semiotics. Toronto: Toronto Semiotic Circle 1976.

Sonesson, Göran, Pictorial concepts. Inquiries into the semiotic heritage and its relevance for the analysis of the visual world. Lund: Lund University Press 1989 (a).

Sonesson, Göran, Semiotics of photography. On tracing the index. Report 4 from the Semiotics project. Lund: Institute of Art History 1989 (b).

Sonesson, Göran, "Icon", in Bouissac, Paul, ed., Encyclopaedia of Semiotics. New York: Garland Publishing Co 1996. (a).

Sonesson, Göran, "Iconicity", in Bouissac, Paul, ed., Encyclopaedia of Semiotics. New York: Garland Publishing Co 1996. (b).

Sonesson, Göran, "Index", in Bouissac, Paul, ed., Encyclopaedia of Semiotics. New York: Garland Publishing Co 1996. (c).

Sonesson, Göran, "Indexicality" in Bouissac, Paul, ed., Encyclopaedia of Semiotics. New York: Garland Publishing Co 1996. (d).

Sonesson, Göran, "The ecological foundations of iconicity", to appear in Semiotics around the world: Synthesis in diversity. Proceedings of the Fifth International Congress of IASS, Berkeley, June 19-18, 1994.

Thom, René, "De l’icône au symbole", in Cahiers internationaux du symbolisme, 22/23, 1973; pp. 85-106.

Williamson, Judith, Decoding Advertisements. London: Marion Boyars 1978.