The Onomasiological Trap

Imposing categories on linquistic data

Jesper Haraldsson

Karlstad University

2024-10-10

Presentation objectives

• define and differentiate between semasiological and onomasiological approaches
• explain the onomasiological trap and its potential impact on linguistic fieldwork
• advocate for methodological pluralism and critical awareness

Nikipini Dalak (left), Glenn Wightman and Lily Bennett recording Dalabon knowledge.

First is to define and differentiate between semasiological and onomasiological approaches. Second that I will explain this trap and its impact on linguistic fieldwork (though it does extend to descirptive linguistics in general). And finally advocate for methodological pluralism and critical awareness. A note on examples I’ll use in this presentation. Given the argument for transitioning away from using introspective (thereby self generated) examples where possible examples presented will be based either on corpora or other scholars. When this is not the case I’ll make it explicitly known.

Image source: Wightman (2019)

Semasiological vs. Onomoasiological Approaches

• Semasiological (Lehmann, 2004; Peters, 2019)
  • Form → Meaning
• Onomasiological (Lehmann, 2004)
  • Meaning → Form

To understand this trap, we need to first distinguish between two fundamental approaches to descriptive linguistics: the semasiological and the onomasiological. The semasiological approach, as its name suggests with sema derived from sign, starts with the forms you observe—the sounds, the words—and then delves into their meanings. For example

The onomasiological approach, on the other hand, flips this script. It begins with a specific meaning or communicative function in mind and then sets out to discover how the language in question expresses that concept. Thereby it offers a seemingly systematic way to uncover the grammatical categories of a language, especially when resources are limited. Methods like elicitation and translation, which we’ll discuss shortly, seem particularly cost-effective and straightforward. 

Unveiling the Trap

both unreliable and invalid […] more a hermeneutic than an empirical method (Lehmann, 2004, p. 24)

The onomasiological approach, particularly in its reliance aforementioned methods, can lead us down a rather treacherous path—one where the data we gather ends up reflecting more about our own expectations as analysts than the true underlying structure of the language we’re studying. In essence, Lehmann cautions us that it is quote “both unreliable and invalid […] more a hermeneutic than an empirical method”. As such if we approach a language with pre-set functional parameters, already looking for how it should express certain meanings based on our own linguistic backgrounds, we risk shaping the data to fit our expectations. The result? We might miss crucial nuances or variations in the language because we weren’t looking for them in the first place. Examples of this to come.

Linguistic elicitation

• Distinction (Majid, 2012)
• Valuable tool (Bowern, 2015; Chelliah, 2001)
• Is a co-construction of meaning (Lehmann, 2004)
• Requires caution and awareness of potential bias (Dale, 1978; Greenbaum & Quirk, 1970; Pickering & Ferreira, 2008; Sampson, 2001; Schütze, 2011)

Now, it’s important not to dismiss the value of linguistic elicitation entirely. And I make this distinction with the premodification because nonlinguistic elicitation has significantly different pros and cons, of which one of the largest pros is the enablement of cross-speaker and cross-language comparisson.

Linguistic elicitation can help us systematically explore the possibilities of a language and reveal categories or structures that might be missed in less controlled settings.

However, Lehmann’s critique reminds us to tread carefully. To illustrate this consider the following self-developed example of a linguist trying to understand how a particular language expresses past events. They might ask an informant, “What did you do yesterday?” intending to elicit past tense forms. However, the specific framing of the question, using “yesterday” as a temporal anchor, might lead the informant to favor certain verb forms over others, potentially obscuring alternative ways the language might express pastness. What if the language has a more nuanced system, distinguishing between recent and distant past events? By imposing our own pre-set category of “yesterday,” we might miss this entirely.

Linguistic elicitation isn’t simply a neutral extraction of data from an informant’s mind. It’s a dynamic, collaborative proces between the linguist and the speaker where meaning is co-constructed. The analyst’s assumptions and background can subtly (or not so subtly) influence the steps taken. The speaker can be susceptible to priming, can reject grammatically accurate statements simply because they violate a prescriptive norm, and after repeated questioning may become more likely to accept previously rejected statements. This potential for bias, for leading the informant rather than truly listening to the language, is what makes critical awareness so crucial.

Interestingly, Schütze’s work offers further support for this need for caution. He notes that even trained linguists, when asked to provide judgments about the grammaticality of sentences, often find that their intuitions differ. If experts can disagree on something as fundamental as whether a sentence is well-formed, the potential for variation and subjectivity when eliciting complex grammatical structures in the field is huge!

Translation

• Interference (Chelliah, 2001; Margetts & Margetts, 2012)
• Obscuration (Evans, 2012)
• Colored judgements (Sampson, 2001)

Similar concerns arise when we consider the translation method. In its simplest form, this method involves taking sentences from a language familiar to the linguist and asking the informant to provide equivalent expressions in the target language. This can be helpful for getting a basic sense of how certain concepts are expressed.

However, a significant challenge, often referred to as “linguistic interference,” emerges when we rely too heavily on translation. Again note the premodification of linguistic as this interference should not be confused with audio or electrical interference when recording language samples. Take for instance the following Lehmann inspired self-developed example, if we ask an informant to translate a sentence containing the English “perfect tense” (e.g., “I have eaten”), we might mistakenly assume that the target language must also have a distinct perfect tense, with its specific nuances of completed action and relevance to the present. However, different languages carve up time and experience in their own unique ways. The target language might have a completely different way of encoding similar meanings, perhaps using aspects or other grammatical devices that don’t map neatly onto English categories. By forcing the translation, we risk imposing a category that doesn’t truly exist in the language, leading to a skewed understanding of its grammar. Furthermore as Chelilah notes the translation may not be idiomatic because of intereference.

Just as our knowledge of our own native language can sometimes lead us to misinterpret sounds or grammatical structures in other languages, so too can it color our judgments and interpretations when relying on translation. We might hear what we expect to hear or see grammatical parallels where none truly exist, simply because our own linguistic system is shaping our perception.

Moving Beyond the Trap

• Corpus analysis (Glynn, 2015; Lehmann, 2004)
• Experimental techniques (e.g., Nielsen, 2015,)

So, where does this leave us? The onomasiological approach, as we’ve seen, can be a powerful tool for uncovering the intricacies of language. But it’s not without its risks. That’s why a cautious, multifaceted approach is essential, one that recognizes the potential for bias and seeks to mitigate it through methodological pluralism. Regarding translation’s role I believe Matthweson expresses it perfectly when they write that tanslations (both of text and elicitations) should quote “be regarded as a clue rather than a result”. What then can be used to gain a result? Lehmann amongst many others argues from the inclusion of corpus analysis on the basis of the quote “enhanced objectivity of the data”. There is also space for the inclusion of a more experimental approach, using techniques like reaction-time measurements, eye-tracking, or even brain imaging can offer another layer of evidence, providing converging support for or challenging findings from more traditional methods. This blending of qualitative and quantitative data, this triangulation of methods, allows us to approach language from multiple angles, strengthening the reliability of our findings.

Quote on translation by Matthewson (2004, p. 380)

Final Thoughts

The most crucial takeaway, the thread that runs through all of these approaches, is the need for critical awareness. We, as linguists, must be constantly aware of our own assumptions, our own linguistic baggage, and how it might color our interpretations. By acknowledging the potential for bias, by remaining open to unexpected findings and alternative explanations, we not only strengthen our research but also honor the languages and communities we study.

References

Bowern, C. (2015). Linguistic Fieldwork. Palgrave Macmillan UK. https://doi.org/10.1057/9781137340801

Chelliah, S. L. (2001). The role of text collection and elicitation in linguistic fieldwork. In P. Newman & M. Ratliff (Eds.), Linguistic Fieldwork (1st ed., pp. 152–165). Cambridge University Press. https://doi.org/10.1017/CBO9780511810206

Dale, I. R. H. (1978). Beyond Intuition: The Use of Questionnaires in Linguistic Investigation. Anthropological Linguistics, 20(4), 158–166. Retrieved from http://www.jstor.org.bibproxy.kau.se/stable/30027472

Evans, N. (2012). Anything can happen : The verb lexicon and interdisciplinary fieldwork. In N. Thieberger (Ed.), The Oxford handbook of linguistic fieldwork (pp. 183–208). Oxford University Press.

Glynn, D. (2015). Semasiology and onomasiology. In J. Daems, E. Zenner, K. Heylen, D. Speelman, & H. Cuyckens (Eds.), Change of paradigms - new paradoxes: recontextualizing language and linguistics. de Gruyter Mouton.

Greenbaum, S., & Quirk, R. (1970). Elicitation Experiments in English: Linguistic Studies in Use and Attitude. University of Miami Press.

Lehmann, C. (2004). Data in linguistics. The Linguistic Review, 21(3-4). https://doi.org/10.1515/tlir.2004.21.3-4.175

Majid, A. (2012). A guide to stimulus-based elicitation for semantic categories. In N. Thieberger (Ed.), The Oxford handbook of linguistic fieldwork (pp. 54–71). Oxford University Press.

Margetts, A., & Margetts, A. (2012). Audio and video recording techniques for linguistic research. In N. Thieberger (Ed.), The Oxford handbook of linguistic fieldwork (pp. 13–53). Oxford University Press.

Matthewson, L. (2004). On the Methodology of Semantic Fieldwork. International Journal of American Linguistics, 70(4), 369–415. https://doi.org/10.1086/429207

Nielsen, L. P. (2015). Knowledge Dissemination Based on Terminological Ontologies: Using Eye Tracking to Further User Interface Design [Phd]. Copenhagen Business School. Retrieved from https://research.cbs.dk/en/publications/knowledge-dissemination-based-on-terminological-ontologies-using-

Peters, P. (2019). Lexicography. In K. Allan (Ed.), The Routledge Handbook of Linguistics (pp. 187–204). ROUTLEDGE.

Pickering, M. J., & Ferreira, V. S. (2008). Structural priming: A critical review. Psychological Bulletin, 134(3), 427–459. https://doi.org/10.1037/0033-2909.134.3.427

Sampson, G. (2001). Empirical linguistics. Continuum.

Schütze, C. T. (2011). Linguistic evidence and grammatical theory. WIREs Cognitive Science, 2(2), 206–221. https://doi.org/10.1002/wcs.102

Wightman, G. (2019). Nikipini Dalak (left), Glenn Wightman and Lily Bennett recording Dalabon knowledge [Photograph]. Retrieved from https://lens.monash.edu/@monash-life/2019/10/27/1377046/preserving-indigenous-languages