Two weeks ago, Dr Lauren Fonteyn, an expert in historical linguistics, wrote an interesting piece on variation in linguistics. To be more precise, she asks what role variation plays in language — does variation just for the sake of variation exist? Linguists, especially linguists of functionalist persuasion, strongly believe that there is no “free” variation: any difference in form spells a difference in meaning. And since Bolinger (1968) convinced the world of this, ever more intricate methods have been adopted in the pursuit of this ideal. Linguists, who long relied on introspection and eschewed quantification, started mining large collections of texts and moved to statistics to pinpoint the differences.

Do the patterns linguists see in language exist in the minds of speakers?

In her answer Dr Fonteyn refers to a (in our view) very exciting question that is fundamental to the Out Of Our Minds research program: do all the patterns that linguists observe in language really exist in the minds of language users?

This is a question that has been asked before and was famously answered by Zellig Harris (1954). Harris said: if you mean by that question whether the patterns we find in the data really are present in the data, then the answer is yes. And with quantitative methods we can easily identify patterns, and decide which ones are flukes. But if you mean by that question whether these patterns exist in the minds of language users, then we should be much more cautious.

And this is an issue that is deserving of much more attention than it is currently getting. We seem to ‘allege [to] a proof’ (Locke 1689/1979): if it exists in the mind of a linguist, it ought to exist in the mind of (each and every) naive user too. Should we not require empirical evidence?

Construal: variation in language meets variation in life

We recently completed a study to investigate whether two expressions that differ ever so slightly from each other affect how we view the situations the are described. We looked at three types of constructions that have been studied extensively in linguistics and/or psycholinguistics.

The bookshelves are above the desk.
The desk is under the bookshelves.

The boy gave the girl a present.
The boy gave a present to the girl.

We let native speakers of English look at still scenes after they had heard one of the two possible descriptions and used eye tracking to capture their gaze. And we found (among many other interesting things) that the effort linguists have put into pinpointing the differences between constructions is negatively correlated with the effect that the differences in linguistic pattern have on perception, and hence conception. That is: the harder linguists have to look to find the differences between constructions, the less these difference affect the average language user. The so-called dative alternation is the prime example: whether you say The boy gave the girl a present or The boy gave a present to the girl barely affects how the scene is viewed.

L’art pour l’art

This does not mean that language users aren’t aware of linguistic variation, nor that it does not affect how they phrase things - there is plenty of research (for the dative alteration, this starts with Bresnan et al 2007 and Bresnan & Ford 2010) that shows that use of the two variants can be reliable predicted in textual data and that language users are sensitive to the triggers that linguists identify.

But even if language users are aware of the contrasts that linguists describe, when and why do they make use of those contrasts? Some variation such as the so-called dative alternation appears to be system internal. Is it “art for art’s sake”?

Some linguistic preferences, although entrenched in memory, seem to remain restricted to the level of the code, indeed. They appear as a ‘larpurlartistic’ device that, essentially, serves to build the plasticity of the communicative system by increasing the expressive potential of the language. A similar idea was recently expressed by De Smet et al (2018) who conclude that we don’t need differences in meaning to motivate variation functionally, because variation itself is functionally motivated.

This does not mean that system internal analyses aren’t valuable. Quite the contrary! Knowing how different options relate to each and to their respective families constitutes important knowledge as it helps us explain and predict language change. And understanding language change better provides new insights into the complex adaptive system that language is.

Whether we as linguists are able to capture the difference between two options or not may ultimately not matter to current language users going about their daily lives. And for linguistics who subscribe to the cognitive commitment, it would be useful to know where to draw the line, before turning into what Lauren Fonteyn calls Ghostbusters.

Dagmar/Petar

References

Bolinger, Dwight. 1968. Entailment and the meaning of structures. Glossa, 2, 119-127.

Bresnan, Joan, Anna Cueni, Tatiana Nikitina, and Harald Baayen. 2007.  "Predicting the Dative Alternation."   In Cognitive Foundations of Interpretation, ed. by G. Boume, I. Kraemer, and J. Zwarts.  Amsterdam: Royal Netherlands Academy of Science, pp. 69-94.

Bresnan, Joan and Marilyn Ford. 2010. "Predicting Syntax: Processing Dative Constructions in American and Australian Varieties of English." Language 86(1): 186-213.

De Smet, Hendrik, Lauren Fonteyn, Frauke D’hoedt and Kristel Van Goethem. 2018. The changing functions of competing forms: Attraction and differentiation. Cognitive Linguistics, 29(2), 197-234.

Fonteyn, Lauren. 2019. Een zoektocht van ‘zo maar’ naar zingeving in taalvariatie. Neerlandistiek: online tijdschrift voor taal-en letterkundig onderzoek. 3 maart 2019.

Harris, Zellig S. 1954. Distributional Structure, WORD, 10:2-3, 146-162.

Locke, J. 1689. An essay concerning human understanding. London: Dent, 1961.