Bodo Winter

Hello Django 😍

I'm definitely in the wrong field

There's a new Collaborative Research Center focusing on the common ground at the University of Tübingen (where I did my MA), with a variety of really cool looking PhD and postdoc opportunities in linguistics starting in October: uni-tuebingen.de/en/283575

Tenure track position in English linguistics (Assistant Professor, Associate Professor or Professor) / tenure track -tehtävä, englannin kielen tutkimus (Apulaisprofessori tai Professori)

We are recruiting! An English linguistics professorship in Tampere (tenure track, open rank). Application deadline August 18
tuni.rekrytointi.com/paikat/?o=A_...

Personal side note: Francesca and I share a love of Ullmann's work...

... and we had so much fun nerding out about this dataset together... for several years now! It's been a fantastic collaboration!

Comments/critique/thoughts are welcome!

But for now, I'm pretty happy how far Francesca and I got with the available data that was already published. ⬇️

There's a lot of limitations which we discuss as well... this work clearly isn't perfect at all, and so much more could be done in the future, in particular, increasing cross-linguistic diversity! Like, we need more data from signed languages, and from non-"WEIRD" spoken languages. ⬇️

Since we also have a pretty exhaustive review of previous work with an extensive bibliography, we hope that this also serves as a great reference for people who want to get into this topic. ⬇️

The paper also makes a lot of points about the interaction between theory/methods in this space.

Besides linguists interested in metaphors, it should be interesting to perceptual psychologists and anybody interested in crossmodal correspondences more generally. ⬇️

In other words: we argue for a shift in the theoretical target, the explanandum, and therefore, we need new explanans.

We argue that like Ullmann, we should be more open-minded about explanations ... not straitjacketing things into a one-size-fits-all principle such as "embodiment". ⬇️

... but really, the literature may have chased a linear phantom, and maybe the explanations were offered too soon.

We say: the thing we seek to explain was actually something different all along, not a sensory cline, but pairwise patterns, and hence we now need different explanations. ⬇️

We don't want to offer explanations for these patterns, and actually caution against jumping to conclusions.

Previously, people tried to explain these tendencies in terms of senses differing in terms of linear differences in "embodiment", "concreteness", "accessibility" etc. ... ⬇️

And if we're honest, the "hierarchy" was never "linear" anyway, and for a while now, representations have been more network-oriented, but people still use verbal characterizations that are at odds with this network view.

This is the coarse summary diagram we come up with on the basis of our data. ⬇️

.. but what we think is a more honest, bottom-up perspective on this data, suggests that we should focus much more on specific, pairwise sensory combinations, as in this key figure below. ⬇️

We argue that several past analyses approaches were very confirmatory in spirit, essentially looking at the data from the top-down perspective of the hierarchy, seeking to re-confirm it in new datasets ... ⬇️

... SPOILER ALERT:

We don't think the notion of the "hierarchy of the senses" holds up, at all.

We use different analyses approaches to look at this complex dataset from different ways — also to trace how others have analyzed this data, and how different methods impact theory building. ⬇️

... but more diverse than any prior sample on synesthetic metaphors. We discuss obvious omissions from this data in the paper.

Either way, using this data, we now assess whether the concept of a hierarchy of the senses holds up (teaser graph of one of our analyses below) ⬇️

Given the structural similarity of these data sets, we're now in a cool position to take stock of the available data and combine it into one analysis.

We found 38 such tables in publications from 1937 to 2022, from 14 distinct languages. Not a typologically balanced sample at all(!!) ... ⬇️

Ullmann published tables like the one below since 1937(!), and as almost all work on synesthetic metaphors follows his foot steps, there's now a host of such tables by other researchers—modern work still publishes the data in this format. ⬇️

... which relates to Western cultural preconceptions about there being a "hierarchy of the senses".

Over time and also through interaction with Viberg's implicational hierarchy of perception verbs, this became "reified" in the form of a strict linear hierarchy. ⬇️

... he observed that in adjective-noun pairs, touch words often fill the adjective slot, and sound words often fill the noun slot—as in "smooth melody" or "rough sound".

He also claimed that intersensory transfers go from the 'lower' to the 'higher' senses ... ⬇️

On the basis of an in-depth qualitative analysis French, Hungarian, and English poetry, the visionary linguist Stephen Ullmann was one of the first to notice that these expressions exhibit striking regularities with respect to which terms can be used to modify which other terms, specifically... ⬇️

Synesthetic metaphors are expressions such as English "smooth melody" or "sweet fragrance" that combine terms from different sensory modalities, in this case, touch/sound and taste/smell. ⬇️

Linguistic Synesthesia Cambridge Core - Cognition - Linguistic Synesthesia

Super proud of my first Cambridge Elements, joint work with my long-term collaborator Francesca Strik-Lievers: "Linguistic Synesthesia: A Meta-Analysis"

www.cambridge.org/core/elements/abs/linguistic-synesthesia/C6B019926C53001D1D698CB0C46F80C6/

Quick summary of the take home message in thread ⬇️

Out now in @pnas.org! 🌹Is a rose by any other name still as roselike?🌹

We study the prevalence of iconicity (does a word look/sound like what it means?) and systematicity (are pronunciation/meaning relationships shared across multiple words?) in large datasets of ASL, English, and Spanish.

🧵1/N

You should try switching to brms! Since going Bayesian with my mixed models, I've been able to say goodbye to lme4 convergence issues and pretty much always fit the model that I wanted.

Ohhh, this looks stellar. Will read asap! :)

Oh yes, and it could also help with research on iconicity more generally of course, not just the r-for-rough pattern. Are you aware of anybody using the database for typological research on iconicity in the lexicon by any chance?

But if I'm honest with you, I'm not sure. On the other hand, if articulation was the main driver and one highlights it, we'd expect iconic phenomena to be strengthened if articulation is highlighted / more relevant, which is not what these studies find. So I'm using it mostly in a disconfirming way.

Ohh, I think that's an interesting idea! The other thing that could be going on is that if — for these phenomena at least — acoustic analogies are the main driver, anything else that detracts from that could weaken the effect...

And it comes together with a reply! It was an interesting and fruitful discussion about the iconicity of different type of r-sounds :)

www.nature.com/articles/s41...