Iconicity in Linguistics Research Paper

At the phonological level, the notion of sound symbolism is such a challenge. Although submorphemic constant correlations between sound and meaning are marginal in languages like English, it has been argued that they are much more pervasive in many languages of Africa and Southeast Asia. At the word level, the corresponding challenge is the notion of onomatapoeia, which has been acknowledged to be marginal in every language. In the artificial registers of poetry, the sound may indeed be made to be an echo to the sense, but this is not the case in flatfooted prose. At the level of compounds and more complex syntagms, however, diagrammatic iconicity in the sense of Peirce (1932) has always been both pervasive and acknowledged: among the staples of linguistic theory which have been strongly iconic are:

(a) The doctrine of semantic compositionality or Saussure’s motivation (Saussure 1916): ‘the meaning of a complex expression is in some way the sum of the meanings of its parts.’

(b) Behaghel’s first law (Behaghel 1932, p. 4): ‘conceptual closeness is reflected in physical closeness.’ One of the clearest demonstrations of this is in Bolinger and Gerstman’s (1957) discussion of the contrasts between phrases like ‘light housekeeper’ and ‘lighthouse keeper,’ in which, as they demonstrated, the perceptual phonetic contrast mirrors the ortho-graphical distinction. Other work in the same vein is Greenberg (1966b)—restatements and elaborations of the traditional insight that derivational affixes are closer to the root than are inflectional ones; Givon (1980)—the degree of embeddedness of a complement clause is a reflection of the degree of control exerted by the subject of the containing clause; and Haiman (1985a)—contrasts between direct and indirect causation and between alienable and inalienable possession, and the semantic consequences of coordination reduction. Related is the idea that the physical incorporation of a string into a larger one corresponds to the degree of its conceptual incorporation in the material into the framework represented by the matrix string. Direct quotation resists incorporation, indirect quotation does not. Relevance conditionals (e.g., ‘If you’re hungry, there’s food in the fridge’) resist incorporation, ordinary conditionals do not in that the former do not ‘count’ as constituents in V 2 languages like German.

(c) ‘The same form is used for the same meaning.’ This is the basis of grammatical analogy (Anttila 1989, p. 89). The somewhat less familiar corollary, that different form is a reflection of different content, is the basis for Breal’s laws of repartition and specialization (Breal 1897, p. 26). The ‘same is same’ heuristic also lies behind all the (never entirely successful) attempts to construct a ‘notional’ definition for parts of speech and grammatical categories in general (e.g., Jakobson 1936, Wierzbicka 1985), as well as for attempts to find explanations for recurrent polysemy of grammatical constructions.

(d) ‘More form reflects more meaning.’ This is still the most widespread diagnostic for markedness (see Greenberg 1966a, pp. 26–7). The clearest demonstration in typological research in Benveniste (1946) on the most frequent expression of the third person singular. As a creative dynamic principle of perception and interpretation, the ‘more is more’ heuristic is identified as what has led to the frequently attested restructuring of verbal paradigms (e.g., Watkins 1962). It is probably also what has led to the excrescence of derivational morphology in some less widely attested cases. In Khmer and in Romanian, for example, a reduced pronunciation of a verb was reinterpreted as the verb, the full form being reinterpreted as a derived nominalization or a derived causative.

(e) The order of clauses (e.g., in ‘I came, I saw, I conquered’) corresponds to the order of events, other things being equal (Jakobson 1965). Elaborations on this ‘post hoc, ergo, post hoc’ heuristic are to be found in the Prague School notion of ‘communicative dynamism,’ itself a derivative of Behaghel’s recognition (1932, p. 4) that the given precedes the new; Greenberg (1966b) on the order of mention of conjuncts, and of the order of protasis and apodosis; and Tai (1985) on the order of constituents within the clause. Hawkins (1994) and others have proposed that Behaghel’s second law (that short constituents precede longer ones, as in Malkiel’s ‘irreversible binomials’ (Malkiel 1959)) may also be reducible to this principle, given the tendency to give reduced expression to what is more familiar or given (frequently old) information.

(f ) Conceptual symmetry is reflected in formal symmetry in a syntagm. Thus coordinate conjunctions, reciprocals, comparisons, and ‘pari passu’ constructions like ‘the more the merrier’ tend to be symmetrically expressed in most languages. More subtly, conceptual symmetry is sometimes expressed by paradigmatic symmetry. The failure of reflexivization in sentences like ‘I like me’ and ‘I want ME to stay behind’ is a reflection of the loss of ‘speaker privilege,’ and hence, indirectly, a direct reflection of a divided self (Haiman 1998, p. 77): ‘me’ is treated no differently from (i.e., symmetrically with) any other participant on the perceptual stage. So, too, conceptual asymmetry may be reflected in formal asymmetry. Langacker (1985) argues that the structure ‘there’s snow all around (sc. me)’ directly reflects the subject’s inability to see himself, and thus reflects the asymmetry between the perceiving subject and the perceived object directly.

In spite of the opacity and arbitrariness in much of language, there is a sense in which we do not believe in our bones that it is real. In his short story ‘Funes the memorious,’ Borges showed what a totally arbitrary code would be:

In place of 7,013, he would say (for example) Maximo Perez; in place of 7,014, The Railroad; other numbers were Luis Melian Lafinur, Olimar, suphur, the reins, the whale, the gas, the cauldron, Napoleon, Agustin de Vedia. In place of 500, he would say nine … (Borges 1963, p. 64). All students of language, however committed they may be to the idea that ‘anything can stand for anything,’ will probably find this conception of a language to be very alien. What Borges’s imaginative and outrageous reductio forces us to acknowledge is the extreme unlikeliness of such willed original opacity. However opaque the signs and combinations of language may seem to us in any mature grammatical system, this system probably owes its present form to an originally iconic code like a pantomime (Donald 1991), modified over the course of time and use by a number of more or less user-friendly transparent signs and rules which ‘seemed to be good ideas at the time.’

Given the acknowledged universality of motivation, analogy, markedness, Behaghel’s law, and the extreme implausibility of truly arbitrary ‘languages’ like the number system of Funes, the interesting question is: what muddies the waters?

One obscuring factor is that limitations of the medium lead to competing (albeit impeccably iconic) motivations. For example, ‘same form’ is motivated not only by ‘same meaning,’ but by symmetry or ‘lack of speaker privilege.’ ‘More form’ is motivated not only by ‘more meaning,’ but also by ‘less frequency’ (Greenberg 1966a, on the relative marking of dual and plural), ‘less familiarity,’ greater surprise (Haiman 1998, on the marking of coreference), and ‘more politeness’ (‘more form’ is ‘more wrapping,’ cf. Geertz 1960). ‘First mention’ is motivated not only by being ‘most familiar,’ but also by Jespersen’s (1949, p. 54) ‘principle of actuality: what is at the moment uppermost in the speaker’s mind tends to be first expressed.’ Overburdened with too many semantic distinctions, a form may come to seem more or less arbitrary (Bolinger 1975, p. 110). So far, the most promising demonstration of the reality of iconicity in spite of this limitation is that where there are no competing motivations for the ‘limited good’ of a particular formal expression, there tends to be agreement (Haiman 1985a on word order).

Opacity or the apparent ‘autonomy of syntax’ is also the result of a number of diachronic factors, of which erosion, routinization, and systematization are perhaps the most important. We can illustrate some of these tendencies in a number system whose iconic origins are still very much in evidence—the Roman numerals.

The first three numbers are of course as iconic as a tally. The numeral V is apparently opaque, but still barely recognizable as a diagram of a human hand. The motivation for the abbreviation was probably no different from the motivation for the sound changes of truncation or assimilation: it is faster and easier to write V than to draw \\\\/. The numeral X can still be seen as a combination of two V’s, one on top of the other. The numbers C and M are also icons, but of a different sort: C (entum) and M(ille) are examples of another kind of abbreviation, and illustrate competing iconic motivations coexisting within the same code.

Implicit in even the simplest compound sign, like II, is a rule which seems so natural that it is hardly recognized at first. It is that the combination XY is to be read as ‘X’+‘Y.’ This is as true of VI, VII, VIII, …, XXXIII, … MMCCCVIII, as it is of II. At some point, an unknown genius recognized and applied analogical reasoning to this natural rule: if XY is read as ‘X’+‘Y,’ then perhaps, under some (very restricted) conditions YX could be read as ‘X’-‘Y.’ The result, apparent in numerals like IV, IX, XC, and the like, is motivated within the system. But these numbers, even when they are drawn in the most painstaking and redundant detail, no longer correspond iconically to the quantities they represent. Analogy has led in this case to something that is the opposite of diagrammatic iconicity. The signs within a system cease to reflect reality and take on a life of their own when they are manipulated by rules internal to the system, as well as by the reality they still attempt to represent.

Within language, aspects of this are recognized as:

(a) the substitution of value for signification within a paradigm (the meaning of signs lies not in their relationship to the world, but in their relationship to each other, cf. Saussure 1916);

(b) grammaticalization, or the routinization of fixed structures beyond the contexts for which they were originally motivated; and

(c) the existence of various kinds of contextsensitivity within a syntagm (Haiman 1998, p. 149).

A compact example of systematization in syntax is the stress pattern of thetic sentences such as ‘TRUMAN died’ (Lambrecht 1988) in languages like English. A stress pattern should iconically reflect the contrast between old and new information, and a word order pattern should do the same. In English, iconic stress has led to sentences such as ‘Truman DIED’ where the predicate is new relative to the subject, but routinization has led to a fixed SVO word order, which is emancipated from the motivation of functional sentence perspective. Given that low stress … HIGH stress means ‘predicate is new,’ perhaps (by an analogical leap similar to the invention of the Roman numeral IV) HIGH stress … low stress means not only ‘subject is new’ (which it still iconically can do), but also ‘everything is new.’

The Roman numeral system is far more iconic than the Arabic system, but it is already fully abstract and digitized. A more primitive system, corresponding to those languages in which there are no numerals higher than some ‘highest digit’ (frequently 2), resorts to subjective and analogic or iconic means of modelling higher numbers, typically one, two, … many . More languages resort to this analogical modelling in the domain of fractions, and present series like one, a half, … a bit . The digitization process ‘colonizes’ the analogic domain, substituting fixed quantities for impressionistic assessments like ‘many’ and ‘a bit.’ A roughly analogous ‘colonization’ process cannot be observed in the lexicon, or in morphosyntax, but it is observed in the conventionalization of intonation (Leon 1993, Fonagy 1977): from being a purely analogical ‘more is more’ system, it becomes grammaticalized and transformed into an ‘all or nothing’ system—although never completely so (Bolinger 1961, 1986).

A future project for evolutionalists is to demonstrate that much of the undisputed arbitrariness of human languages, as of other human institutions, can plausibly be attributed to ritual and system: grammaticalization in the broadest sense.

Bibliography:

1. Anttila R 1989 Historical and Comparative Linguistics, 2nd edn. Benjamins, Amsterdam
2. Behaghel O 1932 Deutsche Syntax, Bd. IV. Carl Winter, Heidelberg
3. Benveniste E 1946 Relations de personne dans le verbe. Bulletin de la Societe de Linguistique 43: 1–12
4. Bolinger D 1961 Generality, Gradience, and the All-or-none. Mouton, The Hague, Netherlands
5. Bolinger D 1975 Aspects of Language, 2nd edn. HBJ, New York
6. Bolinger D 1986 Intonation and its Parts. Stanford University Press, Stanford, CA
7. Bolinger D, Gerstman L 1957 Disjuncture as a clue to constructs. Word 13: 246–55
8. Borges J L 1964 Funes the memorious. In: Labyrinths. New Directions, New York, pp. 59–66
9. Breal M 1897 Essai de semantique. Payot, Paris
10. Donald M 1991 Origins of the Modern Mind. Harvard University Press, Cambridge, MA
11. Fonagy I 1971 Double coding in speech. Semiotica 3: 189–222
12. Geertz C 1960 The Religion of Java. Free Press, Glencoe, IL
13. Givon T 1980 The binding hierarchy and the typology of complements. Studies in Language 4: 333–77
14. Greenberg J 1966a Universals of Language, with Particular Reference to Feature Hierarchies. Mouton, The Hague, Netherlands
15. Greenberg J 1966b Universals of language, with particular reference to the order of meaningful constituents. In: Greenberg J (ed.) 1966 Universals of Language, 2nd edn. MIT Press, Cambridge, MA, pp. 73–113
16. Haiman J 1985a Natural Syntax. Cambridge University Press, Cambridge, UK
17. Haiman J (ed.) 1985b Iconicity in Syntax. Benjamins, Amsterdam
18. Haiman J 1998 Talk is Cheap. Oxford University Press, New York
19. Hawkins J 1994 A Performance Theory of Constituency and Word Order. Cambridge University Press, Cambridge, UK
20. Jakobson R 1936 Beitrag zur allgemeinen Kasuslehre. In: Jakobson R (ed.) Selected Writings, Vol. 2. The Hague, pp. 23–72
21. Jakobson R 1965 Quest for the essence of language. Diogenes 51: 21–37
22. Jespersen O 1949 A Modern English Grammar on Historical Principles, Vol. VII. George Allen and Unwin, London
23. Lambrecht K 1988 Information Structure and Sentence Form. Cambridge University Press, Cambridge, UK
24. Langacker R 1985 Observations and speculations on subjectivity. In: Haiman J (ed.) 1985 Iconicity in Syntax. Benjamins, Amsterdam, pp. 109–49
25. Leon P 1993 Traite de phonostylistique. Nathan, Paris
26. Malkiel Y 1959 Studies in irreversible binomials. Lingua 8: 113–60
27. Peirce C 1932 Philosophical Writings, Vol. 2. Harvard University Press, Cambridge, MA
28. Saussure F de 1916 Cours de linguistique Generale. Payot, Paris
29. Tai J 1985 Temporal sequence and Chinese word order. In: Haiman J (ed.) Iconicity in Syntax. Benjamins, Amsterdam, pp. 49–72
30. Watkins C 1962 Indo-European Origins of the Celtic Verb. Institute of Advanced Study, Dublin, Ireland
31. Wierzbicka A 1985 ‘Oats’ and ‘wheat’. In: Haiman J (ed.) Iconicity in Syntax. Benjamins, Amsterdam, pp. 311–42