All languages manifest two kinds of expressions, prosaic and iconic. In the prosaic layer, meaning–sound correspondences are arbitrary; for example, there is nothing in the string of English sounds d-o-g or Lithuanian š-u-o ‘dog’ that would inherently evoke the image of a devoted furry canine. Conversely, the iconic layer manifests sound–meaning correlations that are not arbitrary. There are two varieties of iconic expressions: onomatopoetic and ideophonic. In both varieties, specific sounds evoke specific meanings, although for different reasons. Onomatopoetic expressions directly imitate sound, as exemplified by the English word p-o-p or the Lithuanian word p-o-k-š-t ‘exploding sound’. In contrast, ideophonic expressions do not directly imitate properties of the external world, but they do lead speakers to unconsciously endow particular sounds with specific sensory meanings. Examples of ideophonic manifestations include front vowels that appear in English words (e.g. teeny-weeny ) associated with small, thin, light things (Jurafsky Reference Jurafsky2014:162–164, among many others), or in Lithuanian diminutive suffixes – yt - or – ėl -.

Ideophones are the focus of this issue. We are drawn to them because linguistically, they remain a mystery. In their form, meaning, and distribution, ideophones usually defy grammar canons specific to a particular language. Therefore, they are often brushed aside as ‘extragrammatical’, or labelled as idiosyncratic and set aside in the designated corner of the lexicon. Nevertheless, data confirming their existence – especially from non-Indo-European languages – have been accumulating for decades (for seminal overviews of such data and the question they raise, see Déchaine Reference Déchaine2015, Dingemanse Reference Dingemanse2012, Hinton et al. Reference Hinton, Nichols and Ohala1994, Nuckolls Reference Nuckolls1999, Smolinsky Reference Smolinsky2001, and Voeltz & Kilian-Hatz Reference Voeltz and Kilian-Hatz2001). And yet, ideophonic expressions continue to elude a satisfactory explanation, regardless of one's theoretical stripes. We hypothesize that this is due to the analytical challenges of forging a model of grammar that would encompass both the prosaic and iconic layers.

This issue grew out of the workshop Structuring Sensory Imagery: Ideophones Across Languages and Cultures, which took place at the University of Rochester on May 2, 2014.Footnote ¹ The workshop itself emerged from our search for unifying threads within the vastly fragmented subfield of sound-symbolic studies (see the aforementioned seminal works) as we wrestled with our own analysis of Finnish ideophones. We purposely invited experts representing diverse underrepresented languages and theoretical backgrounds. That workshop resulted in the papers in the present issue; these papers convey the breadth and depth of questions that linger in sound-symbolic studies. It is our humble hope that we have succeeded in (i) enriching the dialogue by speaking across theoretical divides and different bodies of data; and (ii) conveying the range of issues specific to various distinct approaches.

In their article, Janis Nuckolls, Tod Swanson, Diana Shelton, Alexander Rice and Sarah Hatton (henceforth Nuckolls et al.) invite us to consider the possibility of an audiovisual corpus, which they call an “anti-dictionary,” as a repository for ideophones. They make their case based on an in-depth analysis of representative ideophones and their uses in Pastaza Quichua (Quechuan). Nuckolls et al. argue that conventional dictionaries are designed to communicate abstract, decontextualized meanings; however, while ideophones can communicate fairly abstract meanings, they are not designed to do so. To really understand the nuances of ideophonic meanings, we need to attend to the details of their performance, much as is the case in signed languages such as ASL. When using ideophones, Pastaza Quichua speakers are capable of taking either a speaker-internal perspective (becoming what she or he is depicting) or a speaker-external perspective (portraying a detached observer). Nuckolls et al. argue that failure to capture these nuances is especially consequential in the context of changes within Pastaza Quichuan society, which are leading to an increased dependence on written forms of communication that “erase” much of the malleability of language.

Anthony Webster focuses on Navajo (from the Athabaskan language family) and immerses us in the seductive ideophony of its poetry. This study straddles the boundaries between linguistics, anthropology, literature and literary translation. Language is the glue that keeps the various themes in a delicate (and delectable) balance. Webster takes the view (based on Friedrich Reference Friedrich1986) that poetic language is the locus of the most interesting differences between languages. He follows Sapir (Reference Sapir1921) in assuming that every language can be viewed as a collective art that expresses certain phonetic, rhythmic, and symbolic elements that it does not share with other languages. Webster deftly handles the controversial notion of linguistic determinism, showing that there is a place for it within current linguistic discussions of iconicity.

Webster's study walks us through an example of a particular Navajo poem that is predicated on punning, as is typical of much of Navajo poetry. The discussion revolves around punning as a form of phonological iconicity. Webster reveals the mechanics of Navajo punning, providing extensive notes on its cultural significance. Through a comparison of different versions of the same poem, Webster also addresses the fine craft of translation as a linguistic endeavor and as a social practice. None of the possible translations can be viewed as right or wrong because ambiguity, bivalency, and indeterminacy are intrinsic components of poetry, leaving open the possibility of different renditions. To fully understand a poem and its various renditions, it is necessary to fully comprehend the whole life of a community.

Iraide Ibarretxe invites us to appreciate the robustness of Basque ideophones, lamenting the underestimation and underappreciation of ideophonic studies within mainstream linguistics. She makes her case by first describing the primary characteristics of Basque ideophones and then contextualizing them typologically. In doing so, Ibarretxe provides generous examples and descriptions pertinent to phonetics, phonology, morphology, syntax, semantics, and pragmatics. The typological comparisons are of interest in that she captures not only relevant overlapping properties between Basque (which is an isolate) and other languages, but also the cross-linguistic gaps in patterns that set Basque apart.

Jacob Phillips and David Harrison provide a wealth of examples of ideophonic reduplication in Munda languages (from the Austroasiatic family). They draw on both lexicographic and fieldwork data, remaining as theoretically agnostic as possible and choosing sensory domains as their primary organizing principle in displaying their data. The goal of the paper is to provide a descriptive typology of Munda reduplication which, as they show, spans syntactic categories.

Ludovico Franco studies ideophones in complex predicate constructions. His goal is twofold: to provide a detailed comparative illustration of the phenomenon from a cross-linguistic perspective and to offer a formal account couched within the framework of generative grammar. Franco sets the stage by introducing a number of complex predicate constructions from a range of typologically and structurally unrelated languages. He concludes the empirical section with a summary of possible predicative configurations that include ideophones. Crucially, these are light verb constructions. Next, Franco proposes a syntactic account of these constructions. He adopts a view (from Hale & Keyser Reference Hale, Keyser, Hale and Keyser1993, Reference Hale and Keyser2002, and subsequent work) that syntactic derivations are built compositionally. Based on a battery of tests (distribution, question–answer, number neutrality, use of a particular affix), Franco rules out the possibility that ideophones may be internal arguments of light verbs. Then, broadly following Ramchand (Reference Ramchand2008), he argues that complex predicates with ideophones only ever occur as process or initializing-process event types. Finally, he offers an account of how to build syntactic structures that can host light verb/ideophone predicates.

Solveiga Armoskaite and Päivi Koskinen explore the serialization of nouns in the domain of Finnish (Finno-Ugric) ideophones. They first draw on their corpus data to demonstrate the productivity of a particular type of Finnish noun–noun string, in which an ideophone occurs in one of the nominal slots. They subsequently address the grammatical characteristics of these nominal strings. Finally, building on criteria adapted from verb serialization (Muysken & Veenstra Reference Muysken, Veenstra, Everaert and van Riemsdijk2006), Armoskaite and Koskinen argue that a noun serialization account best captures their Finnish data. They conclude that ideophones, which have a rather marginalized status in the literature, turn out to play a central role in very large theoretical questions about the parallels between verbal and nominal domains.

In the final article of the issue, Kimi Akita takes an experimental approach in exploring the relationship between the morphosyntactic integration of ideophones within a language or a register, and the strength of constraints placed on these ideophones. Presenting data from Japanese, he describes two separate studies. In the first study, Akita shows that sentence-type restrictions, which have been reported for ideophones in several languages and have thus far been associated with a lack of morphological integration, also apply to Japanese ideophones. However, the restriction appears as a preference rather than as an absolute constraint, and is linked to intralinguistic variation in morphosyntactic integration. In his second study Akita investigates the strength of syntactic and semantic restrictions on Japanese ideophonic verbs. His research shows that such restrictions also vary in strength based on morphosyntactic integration, but in this case the variation occurs between different registers of language use.

We conclude with a hope that this issue will stimulate further interest in ideophones. We invite more compilations that (i) take an interdisciplinary approach and (ii) focus on underrepresented languages. We strongly believe that such work will trigger changes in thinking of grammar overall, namely on the relation between the prosaic and iconic layers and how the two interact. One essential avenue of exploration is how to best capture lexicographical aspects of ideophones in their interface with gestures (Nuckolls et al.’s work in this issue is but a start). Another issue that begs for our collective attention is the pragmatic usage of ideophones: while this question comes up frequently (e.g., Armoskaite and Koskinen, Ibarretxe, and Webster, this issue, among many others), we are far from understanding what drives and constrains the usage. Finally, a development of clear methodology for the study of ideophones across distinct grammatical domains and unrelated languages would be a welcome addition to this growing subfield.