Centrum för kognitiv semiotik (CCS)

Humanistiska och teologiska fakulteterna | Lunds universitet

Reconciling iconicity of distance with frequency to explain Paamese grammatical asymmetries

Simon Devylder

The iconicity of distance, or iconicity of cohesion, assumes that “meanings that belong together more closely semantically are expressed by more cohesive forms” (Haspelmath 2008:3). Applied to Paamese, a radical interpretation of the iconicity argument would be that the “cohesive” form natukahin ‘my daughter’ (1) is cohesive (i.e. directly suffixed) because it reflects the close relation between a parent and a daughter. In contrast the distant relation a father-in-law has with his daughter-in-law is reflected in the less cohesive form (i.e. indirectly suffixed) ahineli onak ‘my daughter-in-law’ (2).

    (1)    natu-k-ahin
        ‘my daughter’

    (2)    ahineli ona-k
        female POSS-1SG
        ‘my daughter-in-law’

Haspelmath (2008) firmly opposes the iconicity of distance argument as originally formulated by Haiman (1983) and proposes instead that these grammatical asymmetries are the results of a Zipfian effect of frequency of use leading to shortness and fusion. In other words the form of (1) is “fused” because its (assumed) relative high-frequency of use has lead to fuse the possessive pronoun onak into the possessed noun, compared to the (assumed) low-frequency of the less cohesive form of (2). The author thus rejects that the phenomenon of grammatical asymmetries has anything to do with iconicity, and that instead, his economy-based argument “makes just the right predictions” (2008: 22).
Using the same methodology as Haspelmath (2008) (i.e. a BNCweb corpus-based analysis) I first show that the economy-based argument makes many wrong predictions for Paamese. In line with Croft (2008) I also show the limits of the economy argument based on relative frequency. After showing that the argument does not hold on its own ground, I show the limits of such a methodology that extends the results of an English corpus-based analysis to typologically unrelated languages. Once the limits of the economy argument are identified, I nonetheless support its validity to explain some, but not all, grammatical asymmetries, and thus suggest including it in a more integrated theory of language where economy is considered as one among many forces that compete with each other in giving shape to a language like Paamese. Against the linguistic and anthropological background that I have documented on Paama Island (Devylder 2014), I generally propose that the formal distinction between (1) and (2) is motivated by the multimodal experience of distance in Paamese kinship structure. I precisely show how the topological, functional, sociopragmatic, and affective distance experienced by Paamese islanders become formally captured by kin terms like (1-2) via sociopragmatic frequency of use and diagrammatic iconicity. In sum, I propose to reconcile the traditionally opposed iconicity vs. frequency arguments.


Croft, W. (2008). On Iconicity of Distance. Cognitive Linguistics. Volume 19, Issue 1, Pages 49–57.

Devylder, S. (2014). Paamese Language and Culture (SD1), Digital collection managed by PARADISEC. [Open Access] DOI: 10.4225/72/58ab0479d6eeb

Haiman, J. (1983). Iconic and economic motivation. Language 59, Pages 781–819.

Haspelmath, M. (2008). Frequency vs. iconicity in explaining grammatical asymmetries. Cognitive Linguistics. Volume 19, Issue 1, Pages 1–33.