Sounding out ideas on language, interaction, and iconicity

Mark Dingemanse

Ideophones, like so many things in life, are easy to identify but hard to define. Many researchers have grumbled about the shortcomings of Doke’s descriptive characterization of ideophones (see discussion here), but few have attempted to formulate an alternative. For better or worse, I did,

 1

but it took me a few iterations to arrive at something that I felt worked well enough to be useful in cross-linguistic research.

What’s wrong with “vivid”? The evolution of a definition

One of the benefits of today’s preprint culture is that it is possible to provide constructive critique of pending work before it is out, thereby enabling a rapid cycle of revision before things are committed to print. I have myself benefited from comments on preprints, and have acknowledged such public pre-publication reviews in several of my papers. The below remarks are shared in that spirit.

Pitfalls of fossil-thinking I

Over two years ago I wrote about the unstoppable tide of uninformation that follows the rise of large language models. With ChatGPT and other models bringing large-scale text generation to the masses, I want to register a dystopian prediction.

Monetizing uninformation: a prediction

Writing is thinking. The writing process is the most neglected part of our job. We spend millions on fancy equipment and uncountable hours on training for using this or that toolkit. Yet we assume the BA-level academic writing course we once followed is sufficient; the rest we’ll just learn on the job and hopefully soon we’ll automate away with LLMs. It is all formulaic anyway.

Like anyone active in the space of LLMs and generative AI, I get a growing amount of invitations to speak at conferences and colloquia on something or other LLM-related. A subset of these invites is about the relation between LLMs and academic writing. Recently I had the occasion to respond to such an invite.

Find your own voice: against LLM slop in academic writing

With Times Higher Education writing about citation gaming and hyperprolific authors (surely not unrelated) I hope we can save some of our attention for what Uta Frith and others have called slow science.

Team science is slow science

The closing paragraphs of my previous post were cited in several places (e.g. Culture Making, Far Outliers) as evidence of a cultural revival. Although I feel it is really too soon to say whether this is the case, I’m glad to report that the dirges that we recorded in Akpafu-Todzi are in wide circulation now and are even being played during funerals, to great acclaim.

A cultural revival?

The last time I blindly accepted an invitation to speak was in 2012, when I was invited to an exclusive round table on the future of linguistics at a renowned research institute. As a fresh postdoc I was honoured and bedazzled. When the programme was circulated, I got a friendly email from a colleague asking me how I’d ended up there, and whether I thought the future of linguistics was to be all male.

How to avoid all-male panels (manels)

Today’s dish of expressive vocabulary is particularly tasty. It comes from G|ui, a Khoisan language of Botswana.

 1

To Africanists, expressive words from Khoisan languages are of special interest because Khoisan has been claimed on various occasions to lack ideophones, otherwise thought to be one of those linguistic traits that characterize Africa as a linguistic area (Meeussen 1975:3,

 2

Heine &amp;

Fresh wild melon and meat full of gravy: food texture verbs in G|ui (Khoisan)

The most common use of the term ‘ideophone’ today is as a term for a lexical or grammatical category of words. This use goes back to C.M. Doke, who introduced the term in this sense for the description of Bantu languages. However, Doke did not coin the term. Who did, and how was it used before Doke?

(

 Note:

Looking for a modern definition of the term? Check out “

 ‘Ideophone’ as a comparative concept

” (2019). That chapter supplies the following:

 Ideophone

*. Member of an open lexical class of marked words that depict sensory imagery.* It provides evidence and arguments for the cross-linguistic utility of this definition.) A common term for expressive vocabulary in African linguistics is ‘ideophone’.

On the history of the term ‘ideophone’

Interjections are, in Felix Ameka’s memorable formulation, “the universal yet neglected part of speech” (1992). They are rarely the subject of historical, typological or comparative research in linguistics, and as Aimée Lahaussois has shown (2016), they are notably underrepresented in descriptive grammars. As grammars are the main source of data for typologists, this is of course a perfect example of a self-reinforcing feedback loop.

How should descriptive grammars cover interjections?

A lot of our recent work revolves around working with conversational data, and one thing that’s struck me is that there are no easy ways to create compelling visualizations of conversation as it unfolds over time. The most common form seems to be pixelated screenshots of transcription software not made for this purpose.

Playing with R: unrolling conversation

After much postponement, writing the final report for my NWO Veni grant (2015-2018) turned out to be an unexpected pleasure. It made me realise a couple of things — key among them the role of serendipity in shaping fundamental research. The project was called “Towards a science of linguistic depiction”. Looking at the publications that came out of it and at imperfect indicators like citations, it’s made some useful contributions.

The role of serendipity in shaping fundamental research

DALL-E, a new image generation system by OpenAI, does impressive visualizations of biased datasets. I like how the first example that OpenAI used to present DALL-E to the world is a meme-like koala dunking a baseball leading into an array of old white men — representing at one blow the past and future of representation and generation.

Deep learning, image generation, and the rise of bias automation machines

Some years ago I was following a course by Maarten Kossmann on Tuareg (Tamasheq, Tamajeq, Tamahaq). It was thoroughly enjoyable. After the first lecture we were all alotted a letter of the great

 Dictionnaire Touareg — Français


 1

(a consonant, obviously), and for the remainder of the course these 15 to 40 dictionary pages would form the basis for a number of excercises in the weeks to follow. It was an interesting method;

Expressive verbs in Tuareg

I have been blogging at

 The Ideophone

since 2007, and not all of it has been as ephemeral as my PhD promotor once feared. My short post documenting the etymology of Zotero is apparently the only scientific documentation of where Zotero’s name comes from; it has served as a source in Wikipedia for ages and has received over 15 scholarly citations.

Scholarly blogging, now with DOIs

Large language models make it entirely trivial to generate endless amounts of seemingly plausible text. There’s no need to be cynical to see the virtual inevitability of unending waves of algorithmically tuned AI-generated uninformation: the market forces are in place and they will be relentless.

Large language models and the unstoppable tide of uninformation

PLOS ONE notoriously and astonishingly does not have a proofs stage: authors do not get to see how their work is typeset until the very day it appears. Worse, they have a policy of only correcting formatting errors that they themselves introduce by issuing a formal correction notice — a heavyhanded method that most other journals use only for serious errors of fact or process. Unnecessary corrections hurt authors.

Note: I prepared this posting in August 2015, when PLOS ONE was due to publish a paper by us and I wanted to make sure they’d avoid the stupid typesetting errors they made in our 2013 paper. I used the numbers to convince them to show us proofs beforehand.

Why PLOS ONE needs page proofs

Travel journals provide some of the first written sources on Akpafu. I have previously posted an excerpt from a 1887 journal by David Asante. This here is an excerpt from a similar journey two years later. The whole journey took three months, but this excerpt relates only a trip to two Akpafu towns on 17-18 December 1889.

A visit to Akpafu by Nicolas Clerk, 1889

Sketches, visualizations and other forms of externalizing cognition play a prominent role in the work of just about any scientist. It’s why we love using blackboards, whiteboards, notebooks and scraps of paper. Many folks who had the privilege of working the late Pieter Muysken fondly remember his habit of grabbing any old piece of paper that came to hand, scribbling while talking, then handing it over to you.

Thinking visually with Remarkable

The oldest written fragments of Siwu found so far come from Rudolph Plehn (1898).

 1

Besides some words and phrases (edited and published in 1899 by his friend Seidel), Plehn took down two lines of songs. To one of them I devoted a post some time ago.

Bingo! Refinding the oldest specimen of Siwu

There is a minor industry in speech science and NLP devoted to detecting and removing disfluencies. In some of our recent work we’re showing that treating talk as sanitised text can adversely impact voice user interfaces. However, this is still a minority position. Googlers Dan Walker and Dan Liebling represent the mainstream view well in this blog post:  Fair enough, you might say.

Sawing off the branch you’re sitting on

Language makes us human. But there is an interesting asymmetry in our willingness to ascribe linguistic capacities to non-humans: animals are seen as having none, whereas computers may well master it according to many. What curious conception of language makes this asymmetry possible? And what do Descartes and Turing have to do with it? Notes from a new essay about language between animals and computers.

Language is what makes us human: one of those things, perhaps

 the

one thing, that sets us apart. But there is an interesting asymmetry in our willingness to ascribe linguistic capacities to non-humans: animals tend to be seen as having none, whereas computers are increasingly thought to have mastered language. This asymmetry is the focus of a recent essay I co-authored with a range of people, led by Marlou Rasenberg.

Language between animals and computers

Few historical maps of Ghana’s Volta and Oti regions have been invested with so much political and sociohistorical meaning as Hans Gruner’s 1913 map of the Togo Plateau. Gruner, stationed for over twenty years at Misahöhe in present-day Togo, was a long-time colonial administrator known for his ethnographical and historical knowledge of the area.

The Gruner Map: a 1913 map of the Togo Plateau in present-day Ghana

This is a the second part in a two part series of peer commentary on a recent preprint. The first part is here. I ended that post by noting I wasn’t sure all preprint authors were aware of the public nature of the preprint. I am now assured they are, and have heard from the senior author that they are working on a revised version.

Pitfalls of fossil-thinking part II

Clark &amp; Fischer propose that people see social robots as interactive depictions and that this explains some aspects of people’s behaviour towards them. We agree with C&amp;F’s conclusion that we don’t need a novel ontological category for these social artefacts and that they can be seen as intersecting with a lineage of depictions from Michelangelo’s David to Mattel’s talking barbie doll. We have two constructive contributions to make.

How robots become social: A comment on Clark &amp; Fischer

The construction of gothic cathedrals like Chartres was governed not by blueprints but by “talk, tradition, and templates” — at least that is what Turnbull has compellingly argued. When you come across such a neatly alliterative triad, there are two ways you can go. You can adopt the terms in an unexamined way and rely on their alliterative power.

Reading Suchman’s classic

 Human-machine reconfigurations: plans and situated actions

, I am impressed by her description David Turnbull’s work on the construction of gothic cathedrals. In brief, the intriguing point is that no blueprints or technical drawings or even sketches are known to have existed for any of the early modern gothic cathedrals, like that of Chartres.

Talk, tradition, templates: a meta-note on building scientific arguments

One of the key tasks scientists need to master is how to manage bibliographic information: collecting relevant literature, building a digital library, and handling citations and bibliographies during writing. This tutorial introduces Zotero (www.zotero.org), an easy to use reference management tool made by scholars for scholars. The tutorial covers the basics of using Zotero for collecting, organizing, citing and sharing research.

Slides for a hands-on Zotero workshop

A preprint claims that “ideas from theoretical linguistics have played no role in [NLP]”. Outside the confines of Chomskyan linguistics folks have long been working on  incorporating storage, retrieval, gating and attention in theories of language, with direct relevance to computational models. The only way to give any content to the claim is by giving the notion “theoretical linguistics” the narrowest conceivable reading.

This Lingbuzz preprint by Baroni is a nice read if you’re interested in linguistically oriented deep net analysis. I did feel it’s a bit hampered by the near-exclusive equation of linguistic theory with generative/Chomskyan aps.

Linguistic roots of connectionism

There is a considerable halo-effect attached to JIFs, whereby an article that ends up in a high IF journal (whether by sheer brilliance or simply knowing the right editor, or both) is treated, unread, with a level of veneration normally reserved for Wunderkinder. Usually this is done by people totally oblivious to network effects, gatekeeping and institutional biases.

I’ve been caught up in a few debates recently about

 Recognition and Rewards

, a series of initiatives in the Netherlands to diversify the ways in which we recognize and reward talent in academia. One flashpoint was the publication of an open letter signed by ~170 senior scientists (mostly from medical and engineering professions), itself written in response to two developments.

Why article-level metrics are better than JIF if you value talent over privilege

Wikidata is an ambitious enterprise, but social ontologies are never language-agnostic — so the project risks perpetuating rather than transcending the worldviews most prevalent in current Wikipedia databases, which means broadly speaking global north, Anglo, western, white cishet male worldviews. I think Wikidata is perhaps promising for brute physical facts like the periodic table and biochemistry.

TL;DR

Wikidata is an ambitious enterprise, but social ontologies are never language-agnostic — so the project risks perpetuating rather than transcending the worldviews most prevalent in current Wikipedia databases, which means broadly speaking global north, Anglo, western, white cishet male worldviews. I think Wikidata is perhaps promising for brute physical facts like the periodic table and biochemistry.

Concrete reasons to be skeptical about an ‘Abstract Wikipedia’

I note with sadness that William J. Samarin has passed away in Toronto on January 16, 2020 at the age of 93. An all too short obituary notes that he was “known for his work on the language of religion and on two Central African languages: Sango and Gbeya”.  In linguistics, Samarin was of course also known for his extensive work on ideophones, playful and evocative words with sensory meanings.

Farewell, Mr. Ideophone: William J. Samarin (1926-2020)

Academia.edu takes your academic work and puts it behind a privacy-defying signup form, laces it with ads, botches the metadata, and tries to appease you by offering stats and social ranking that promote constant comparison. What’s not to like?

A while back some low quality citations started showing up on Google Scholar. They had titles like “CHAPTER 2 draft — email xyz@ab.edu” and it was hard find actual bibliographic metadata. Google Scholar seemed to have scraped random PDFs uploaded on Academia.edu and decided it was worth counting the citations in them even in the absence of proper metadata.

How Academia.edu promotes poor metadata and plays to our vanity

I have always had a fondness for things considered marginal in linguistics. The tide may be turning for at least some marginalia: work on ideophones is clearly on the rise, and initiatives such as Martina Wiltschko’s Eh lab at UBC and a new nonlexical vocalizations project at Linköping University show there is significant interest in this area.

New!

Some of this is now published here (open access, free for all!): Dingemanse, Mark. 2020.

What is ‘non-lexical’? Notes on non-lexical vocalisations, II

I have always had a fondness for things considered marginal in linguistics. The tide may be turning for at least some marginalia: work on ideophones is clearly on the rise, and initiatives such as Martina Wiltschko’s Eh lab at UBC and a new nonlexical vocalizations project at Linköping University show there is significant interest in this area. Part I from my notes on a workshop on ‘Ideophones and non-lexical vocalizations’.

Last week I was happy to present my work at a workshop on

 Ideophones and nonlexical vocalisations

in Linköping, Sweden, organised by Leelo Keevallik and Emily Hofstetter. This was the kick-off for a new project on “Non-lexical vocalisations”. It was my first time in Linköping and it was great getting to know the vibrant community of interaction researchers from across departments.

A variety of vocal depictions: Notes on non-lexical vocalisations, I

This is the first ever published account of a visit to Akpafu. It was written down by David Asante, a Twi pastor who travelled throughout today’s Volta Region in the company of some white missionaries. The journey took place in January 1887; the date of the visit to Akpafu was January 25th, 1887.

A visit to Akpafu by David Asante, 1887

Words evolve not as blobs of ink on paper but in face to face interaction. The nature of language as fundamentally interactive and multimodal is shown by the study of ideophones, vivid sensory words that thrive in conversations around the world. The ways in which these “Lautbilder” enable precise communication about sensory knowledge has now for the first time been studied in detail.

Words   Few professions should be more familiar with the nature of words than academia. Words are the currency of our trade. They record our cumulative progress and they measure our productivity as we disperse our ideas through articles and books. How easy is it to fall in love with the printed word, black symbols on a white page, tidy spaces separating units of thought like stars dotting the skies of conceptual clarity!

How to paint with language

It’s easy to forget amidst a rising tide of synthetic text, but language is not actually about strings of words, and language scientists would do well not to chain themselves to models that presume so. For apt and timely commentary we turn to Bronislaw Malinowski who wrote:  In follow-up work, Malinowski has critiqued the unexamined use of decontextualised strings of words as a proxy for Meaning:  Malinowski did not write this on his substack,

Malinowski (1922) on Large Language Models

Lezenswaardig: een groep jonge medici ageert tegen de marketing-wedstrijd waarin volgens hen narratieve CVs in kunnen ontaarden — de nieuwste bijdrage aan het Erkennen &amp; Waarderen-debat. Maar niets is wat het lijkt. Over evidence-based CVs, kwaliteit &amp; kwantificatie.

Eerst dit: de brief benoemt het risico dat je met narratieve CVs een soort competitie krijgt tussen verhalen.

Van betekenisloze getallen naar een evidence-based CV

I have other things to do but one day I’ll enlarge on the insidious effects of elevating this cursed little histogram of “Research output per year” as the single most important bit of information about academics at thousands of universities that use Elsevier Pure. Consider this mini-rant my notes for that occasion. Most importantly, we DO NOT write per year. Our careers are too diverse and precarious to measure output that way.

A rant about Elsevier Pure

Will synthetic text generators usher in a new age of creative thinking? The remarkable fluency of large language models may make them interesting tools for rapidly exploring semantic and stylistic spaces, yet the deceptive ease with which they generate output also provides countless new ways of appropriating ideas and erasing authorship.

Human creativity in the age of text generators

Magritte’s best known work by far is of course his drawing of a pipe with the text

 Ceci n’est pas une pipe

. He made several versions over the years, but the work originated in 1928 or 1929. The title Magritte gave to this painting is

 La trahison des images

— the treachery of images.

Magritte on Words and Images (PDF)

Having been a small and quite isolated language for centuries, Siwu was relatively late to attract attention from outsiders. Europeans in search for gold to buy and people to enslave for the most part stayed near the coast.

‘Remnants of some ancient tribal idiom’: deciphering the oldest Siwu to appear in print

One of my projects here at The Ideophone has been to track down early sources on ideophonic phenomena. For example, I have suggested that we may call the 1850’s the decade of the discovery of ideophones in African linguistics. But we can push back the linguistic discovery of ideophones a little further by looking to other traditions.

Aduerbia sonus: Ideophones in two 17th century grammars of Japanese

Zotero takes it name from an Albanian verb minus some morphology.

If you’ve read yesterday’s post (Zotero, an Endnote alternative) or come across Zotero elsewhere, you may have been wondering about its name.

The etymology of Zotero

Featured philologist of today is Joh. Bernhard Schlegel, for providing us with precious data on ideophones (expressives) in nineteenth-century Ewe, a Kwa language of southeastern Ghana.

This is the first post in a series. Featured philologist of today is Joh. Bernhard Schlegel, for providing us with precious data on ideophones (expressives) in nineteenth-century Ewe, a Kwa language of southeastern Ghana. But since this is the first post on ideophones here, let’s first try to answer the obvious question: what are ideophones, anyway?

Early sources on African ideophones, part I: Schlegel on Ewe, 1857

Sigismund Wilhelm Koelle is one of the founding fathers of African linguistics, and 1854 was one of his more productive years: he published the first large-scale comparison of some 200 African languages (the famed Polyglotta Africana), but also a corpus of Kanuri folklore, a grammar of Vai, and a grammar of Kanuri.

It is high time for a continuation of our series honouring the ancestors of ideophone studies. Sigismund Wilhelm Koelle is one of the founding fathers of African linguistics, and 1854 was one of his more productive years.

Early sources on African ideophones, part IV: S.W. Koelle on Kanuri, 1854

We don’t generally see PhD dissertations as an exciting genre to read, and that is wholly our loss. As the publishing landscape of academia is fast being homogenised, the thesis is one of the last places where we have a chance to see the unalloyed brilliance of up and coming researchers. Let me show you using three examples of remarkable theses I have come across in the past years.

Mindblowing dissertations

Google Scholar is useful, but its inclusiveness and mix of automatically updated and hand-curated profiles means you should never take any of its numbers at face value. Case in point: the power couple Prof.

*

 Summary:

Google Scholar is useful, but its inclusiveness and mix of automatically updated and hand-curated profiles means you should never take any of its numbers at face value. Case in point: the power couple Prof.

Some things you need to know about Google Scholar

Knots are fascinating: they tie together topology, embodied experience, and material culture. As Thai textile artist and designer Nithikul Nimkulrath (2024) has pointed out, knots are the kind of thing we come to know “

 through

and

 in

making”. One of her artworks (see photo) explorers the materiality of knots by translating a physical, hand-knotted container into a 3D render.

Tying ourselves into knots about reasoning

Funeral dirges (

 sìnɔ

in Siwu) are sung during the period of public mourning preceding a burial. The musical structures of these dirges, the performances, and their place in the larger context of the funeral have been described in some detail by Agawu (1988) and before him by the German missionary Friedrich Kruse (1911); however, the linguistic aspects of the genre have not received any attention so far.

I thought I had company (a Mawu dirge)

The Ideophone

DALL-E, a new image generation system by OpenAI, does impressive
visualizations of biased datasets. I like how the first example that
OpenAI used to present DALL-E to the world is a meme-like koala dunking
a baseball leading into an array of old white men --- representing at
one blow the past and future of representation and generation.

<figure
class="wp-block-embed is-type-rich is-provider-twitter wp-block-embed-twitter">
<div class="wp-block-embed__wrapper">
<blockquote>
Our newest system DALL·E 2 can create realistic images and art from a
description in natural language. See it here: <a
href="https://t.co/Kmjko82YO5">https://t.co/Kmjko82YO5</a> <a
href="https://t.co/QEh9kWUE8A">pic.twitter.com/QEh9kWUE8A</a>
— OpenAI (@OpenAI) <a
href="https://twitter.com/OpenAI/status/1511707245536428034?ref_src=twsrc%5Etfw">April
6, 2022</a>
</blockquote>
</div>
</figure>

It's easy to be impressed by cherry-picked examples of DALL•E 2 output,
but if the training data is web-scraped image+text data (of course it
is) the ethical questions and consequences should command much more of
our attention, [as argued
here](https://arxiv.org/abs/2110.01963){rel="noreferrer noopener"
target="_blank"} by Abeba Birhane and Vinay Uday Prabhu.

Suave imagery makes it easy to miss what
[\#dalle2](https://twitter.com/hashtag/dalle2) really excels at:
automating bias. Consider what DALL•E 2 produces for the prompt "a data
scientist creating artificial general intelligence":

<figure
class="wp-block-embed is-type-rich is-provider-twitter wp-block-embed-twitter">
<div class="wp-block-embed__wrapper">
<blockquote>
<a
href="https://twitter.com/hashtag/dalle?src=hash&amp;ref_src=twsrc%5Etfw">#dalle</a>
<a href="https://t.co/IdieXIYWlu">pic.twitter.com/IdieXIYWlu</a>
— Boris Power (@BorisMPower) <a
href="https://twitter.com/BorisMPower/status/1512266548743917599?ref_src=twsrc%5Etfw">April
8, 2022</a>
</blockquote>
</div>
</figure>

::: {.wp-block-image}
<figure class="alignright is-resized">
<img src="https://pbs.twimg.com/media/FP-FAgLWQAIU74_.jpg"
style="width:-782px;height:-762px" decoding="async" />
</figure>
:::

When the male bias was pointed out to AI lead developer Boris Power, he
countered that "it generates a woman if you ask for a woman". Ah yes,
what more could we ask for? The irony is so thicc on this one that we
should be happy to have ample
[\#dalle2](https://twitter.com/hashtag/dalle2) generated techbros to
roll eyes at. It inspired me to make a meme. Feel free to use this meme
to express your utter delight at the dexterousness of DALL-E, cream of
the crop of image generation!

## The systematic erasure of human labour {#the-systematic-erasure-of-human-labour .wp-block-heading}

It is not surprising that glamour magazines like Cosmopolitan,
self-appointed suppliers of suave imagery, are the first to fall for the
gimmicks of image generation. As its editor Karen Cheng found out after
thousands of tries, it generates a woman if you ask for "a female
astronaut with an athletic feminine body walking with swagger" (Figure
3).

I also love this triptych because of the evidence of human curation in
the editor's tweet ("after thousands of options, none felt quite
right...") --- and the glib erasure of exactly that curation in the
subtitle of the magazine cover: "and it only took 20 seconds to make".

The erasure of human labour holds for just about every stage of the
processing-to-production pipeline of today's image generation models:
from data collection to output curation. Believing in the magic of AI
can only happen because of this systematic erasure.

<figure class="wp-block-image size-full">
<a href="https://ideophone.org/files/image-37.png"><img
src="https://ideophone.org/files/image-37.png" class="wp-image-5688"
loading="lazy" decoding="async"
srcset="https://ideophone.org/files/image-37.png 718w, https://ideophone.org/files/image-37-300x174.png 300w, https://ideophone.org/files/image-37-500x290.png 500w"
sizes="(max-width: 718px) 100vw, 718px" width="718" height="417" /></a>
<figcaption>Figure 3</figcaption>
</figure>

*Lezenswaardig: een [groep jonge medici ageert
tegen](https://www.scienceguide.nl/2021/07/nieuwe-erkennen-en-waarderen-kan-kwetsbare-groepen-schaden/){rel="noreferrer noopener"
target="_blank"} de marketing-wedstrijd waarin volgens hen narratieve
CVs in kunnen ontaarden --- de nieuwste bijdrage aan het Erkennen &
Waarderen-debat. Maar niets is wat het lijkt. Over evidence-based CVs,
kwaliteit & kwantificatie.*

Eerst dit: de brief benoemt het risico dat je met narratieve CVs een
soort competitie krijgt tussen verhalen. Dat kan zeker als de conventies
van het genre nog niet uitgekristalliseerd zijn, zoals ik al [schreef in
2019](https://ideophone.org/narrative-cv-as-an-experiment-in-convergent-cultural-evolution/){data-type="post"
data-id="5984" target="_blank" rel="noreferrer noopener"}, toen NWO het
invoerde. Een mooie-verhalen-wedstrijd wil niemand, daar zijn we het
over eens. Ik ben het wat dat betreft trouwens ook eens met misschien
wel het belangrijkste punt van de eerste brief o.l.v. Raymond Poot:
meten is weten. Je moet alleen wel weten wát je meet. Daarover gaat dit
stuk ook.

De medici (zowel deze jongere collega's als de senioren o.l.v.
[@raymondpoot](https://twitter.com/raymondpoot) in het openingssalvo)
lijken vooral te ageren tegen de term "narratief CV". Die heeft ook de
schijn tegen natuurlijk: gaan we elkaar nou sterke verhalen zitten
vertellen bij het kampvuur? Nee toch zeker! Volgens de briefschrijvers
moet een wetenschapper in het nieuwe systeem iets over haar achtergrond
& prestaties opschrijven "op een onderscheidende manier" en "zonder
kwantitatieve maten te gebruiken". Factcheck: ❌ Kwantificatie in het
narratieve CV is prima, gewenst zelfs!

Laten we de call van NWO er anders even bij pakken: [hier is de
PDF](https://www.nwo.nl/sites/nwo/files/media-files/Call%20for%20Proposals%20Vidi%202021%20NL_2.pdf){rel="noreferrer noopener"
target="_blank"} --- het stuk waar het om gaat (§3.4.1 sectie 1 en 2)
plak ik hieronder

<figure class="wp-block-image size-full">
<a href="https://ideophone.org/files/image-64.png"><img
src="https://ideophone.org/files/image-64.png" class="wp-image-5990"
loading="lazy" decoding="async"
srcset="https://ideophone.org/files/image-64.png 597w, https://ideophone.org/files/image-64-199x300.png 199w"
sizes="(max-width: 597px) 100vw, 597px" width="597" height="900" /></a>
</figure>

## Evidence-based CV {#evidence-based-cv .wp-block-heading}

Als de term "narratief CV" je niet zint kun je het ook een
**evidence-based CV** noemen: in plaats van contextloze lijstjes &
getallen wil men argumenten zien voor de excellentie van de kandidaat &
haar werk, kracht bijgezet door kwalitatief en kwantitatief bewijs van
impact.

Want kijk even mee: zowel *kwantitatieve* als *kwalitatieve* indicatoren
zijn uitdrukkelijk toegestaan. Dat zou je niet uit de brieven van
Raymond Poot & collega-medici gehaald hebben. Het cruciale verschil is
dat indicatoren duidelijk betrekking moeten hebben op specieke items:
"Alle type kwaliteitsindicatoren mogen genoemd worden, zolang ze
betrekking hebben op slechts één output item."

Wat hier goed aan is 1: Waar eerder de complete publicatielijst geplempt
mocht worden (waar vooral veelschrijvers bij gebaat zijn) vraagt dit
format om een gemotiveerde keuze van 10 items: de [n-best
methode](http://doi.org/10.1016/j.tics.2019.09.010){target="_blank"
rel="noreferrer noopener"} die aan Ivy Leagues gangbaar is. Niks mis
mee!

Wat hier goed aan is 2: Waar je eerder goede sier kon maken met
journal-level metrics als IF (statistisch gezien niet meer dan een
opgedirkt halo-effect) moet je nu hard bewijs leveren voor de impact &
het belang van je werk.

Wat hier goed aan is 3: Waar je eerder te koop kon lopen met een hoge
h-index (niet gecorrigeerd voor [voorsprong door
leeftijd](http://doi.org/10.1002/asi.22976){rel="noreferrer noopener"
target="_blank"}, coauteurschap, zelfcitaties & andere biases) mag je nu
laten zien welke van je papers echt zo briljant & origineel zijn.

## Dat kwantificatie niet meer mag is quatsch[]() {#dat-kwantificatie-niet-meer-mag-is-quatsch .wp-block-heading}

Volgens mij zijn dat ook 3 manieren waarop een evidence-based CV meer
kansen biedt juist voor de 'kwetsbare groepen' die de brief noemt. (En
ook: 3 manieren waarop de voorsprong van traditioneel bevoorrechten
enigszins rechtgetrokken wordt --- is dat niet ook een deel van de
pijn?)

Kortom, dat kwantificatie niet meer zou mogen is quatsch. Je kunt alleen
niet meer wegkomen met de meest indirecte cijfers (die vooral wat zeggen
over privileges, kruiwagens en co-auteurs) --- in plaats daarvan moet je
nu hard bewijs leveren voor de impact & het belang van je werk.

Ik moet wel zeggen: de misverstanden in de brieven komen niet helemaal
uit de lucht vallen. "Narratief CV" is geen beste term en er is
kennelijk gebrek aan sterke voorbeelden van verantwoorde & genuanceerde
kwantificatie op artikelniveau. Werk aan de winkel voor Erkennen en
Waarderen en NWO!

Tot slot: álle briefschrijvers ---van
[@raymondpoot](https://twitter.com/raymondpoot) cs tot
[@DeJongeAkademie](https://twitter.com/DeJongeAkademie)
[@RadboudYA](https://twitter.com/RadboudYA) etc tot de jonge medici---
zijn het erover eens dat roofbouw op de financiering de echte nekslag is
voor topwetenschap in ons land: meer investering in fundamenteel
onderzoek is cruciaal

*Toevoeging 11 mei 2022:*

Nou, mijn betoog in dit draadje, of in ieder geval de de term
'evidence-based CV', lijkt bij NWO gehoor gevonden te hebben --- waar op
mijn CV zal ik dat zetten? 😃

<figure class="wp-block-image">
<img src="https://pbs.twimg.com/media/FSdS_xtXsAAxe4n.jpg"
decoding="async" />
</figure>

*Originally tweeted by
[@dingemansemark](https://scholar.social/@dingemansemark){.u-url
.mention rel="mention"}
([@DingemanseMark](https://twitter.com/DingemanseMark)) on [July 27,
2021](https://twitter.com/DingemanseMark/status/1420124953014329350).*

This [Lingbuzz preprint by
Baroni](https://ling.auf.net/lingbuzz/006031){rel="noreferrer noopener"
target="_blank"} is a nice read if you're interested in linguistically
oriented deep net analysis. I did feel it's a bit hampered by the
near-exclusive equation of linguistic theory with generative/Chomskyan
aps. (I know it makes a point of claiming a "very broad notion of
theoretical linguistics", but it doesn't really demonstrate this, and
throughout the implicit notion of theory is near-exclusively aligned
with GG and its associated concerns of competence, poverty of the
stimulus, et cetera).

<figure class="wp-block-image">
<img src="https://pbs.twimg.com/media/E4FDxONXwAMFvCh.png"
decoding="async" />
</figure>

For instance, it notes (citing Lappin) that theoretical linguistics
"played no role" in deep learning for NLP, but while this may hold for
generative grammar (GG), linguistic theorizing was much broader than
that right at the start of connectionism and RNNs, e.g. in [Elman
1991](http://doi.org/10.1023/A:1022699029236){target="_blank"
rel="noreferrer noopener"}.

In fact, just look at the bibliography of Elman's classic RNN work and
tell us again how exactly theoretical linguistics "played no role" ---
Bates & Macwhinney, Chomsky, Fillmore, Fodor, Givon, Hopper & Thompson,
Lakoff, Langacker, they're all there. Elman's bibliography is a virtual
*Who is Who* of big tent linguistics at the start of the 1990s. The only
way to give any content to Lappin's claim (and by extension, Baroni's
generalization) is to give the notion of "theoretical linguistics" the
narrowest conceivable reading.

<figure
class="wp-block-gallery has-nested-images columns-default is-cropped wp-block-gallery-3 is-layout-flex wp-block-gallery-is-layout-flex">
<figure class="wp-block-image size-large">
<a href="https://ideophone.org/files/E4FEkLuWUAI6IwO.png"><img
src="https://ideophone.org/files/E4FEkLuWUAI6IwO-696x1024.png"
class="wp-image-5640" loading="lazy" decoding="async" data-id="5640"
srcset="https://ideophone.org/files/E4FEkLuWUAI6IwO-696x1024.png 696w, https://ideophone.org/files/E4FEkLuWUAI6IwO-204x300.png 204w, https://ideophone.org/files/E4FEkLuWUAI6IwO.png 754w"
sizes="(max-width: 696px) 100vw, 696px" width="696" height="1024" /></a>
</figure>
</figure>

However, Baroni's point may generalize: perhaps modern-day usage-based,
functional, and cognitive approaches to ling theory aren't drawing as
heavily on current NLP/ML/DL work as they could either. Might a lack of
reciprocity play a role? After all, the well known ahistoricism and lack
of interdisciplinary engagement of NLP today does not exactly invite
productive exchange. (Though some of us try.)

The theory=Chomsky equation also makes it appearance at the end, where
Baroni muses about incorporating storage, retrieval, gating and
attention in theories of language. Outside the confines of Chomskyan
linguistics folks have long been working on precisely such things. One
might think work by Joan Bybee, Maryellen MacDonald, Morten
Christiansen, and others might merit a mention!

<figure class="wp-block-image">
<img src="https://pbs.twimg.com/media/E4FH0LUXwAY9NfF.png"
decoding="async" />
</figure>

In sum, Baroni's piece provides an informative if partial review of
recent work and includes bold proposals (e.g., deep nets as algorithmic
linguistic theories), worth reading if you're interested in a particular
kind of linguistics. Consider pairing it with this [well-aged bottle of
Elman 1991](http://doi.org/10.1023/A:1022699029236){target="_blank"
rel="noreferrer noopener"}!

## References {#references .wp-block-heading}

- Baroni, M. (2021, June). *On the proper role of
 linguistically-oriented deep net analysis in linguistic theorizing*.
 LingBuzz. Retrieved from <https://ling.auf.net/lingbuzz/006031>
- Bybee, J. L. (2010). *Language, Usage, and Cognition*. Cambridge:
 Cambridge University Press.
- Christiansen, M. H., & Chater, N. (2017). Towards an integrated
 science of language. *Nature Human Behaviour*, *1*,
 s41562-017-0163--017. doi:
 [10.1038/s41562-017-0163](https://doi.org/10.1038/s41562-017-0163)
- Elman, J. L. (1991). Distributed Representations, Simple Recurrent
 Networks, And Grammatical Structure. *Machine Learning*, *7*,
 195--225. doi:
 [10.1023/A:1022699029236](https://doi.org/10.1023/A:1022699029236)
- Lappin, S. (2021). *Deep learning and linguistic representation*. Boca
 Raton: CRC Press.
- MacDonald, M. C., & Christiansen, M. H. (2002). Reassessing working
 memory: Comment on Just and Carpenter (1992) and Waters and Caplan
 (1996). *Psychological Review*, *109*(1), 35--54. doi:
 [10.1037/0033-295X.109.1.35](https://doi.org/10.1037/0033-295X.109.1.35)

*Originally tweeted by
[@dingemansemark](https://scholar.social/@dingemansemark){.u-url
.mention rel="mention"}
([@DingemanseMark](https://twitter.com/DingemanseMark)) on [June 17,
2021](https://twitter.com/DingemanseMark/status/1405488247057391616).*

::: {.wp-block-image}
<figure class="alignright size-full is-resized">
<a href="https://ideophone.org/files/image-22.png"><img
src="https://ideophone.org/files/image-22.png" class="wp-image-5618"
style="width:360px;height:206px" loading="lazy" decoding="async"
srcset="https://ideophone.org/files/image-22.png 947w, https://ideophone.org/files/image-22-300x171.png 300w, https://ideophone.org/files/image-22-768x438.png 768w, https://ideophone.org/files/image-22-500x285.png 500w"
sizes="(max-width: 947px) 100vw, 947px" width="947" height="540" /></a>
</figure>
:::

I have other things to do but one day I'll enlarge on the insidious
effects of elevating this cursed little histogram of "Research output
per year" as the single most important bit of information about
academics at thousands of universities that use Elsevier Pure. Consider
this mini-rant my notes for that occasion.

Most importantly, we DO NOT write per year. Our careers are too diverse
and precarious to measure output that way. What we write is key, where
to find it is important, but how much per calendar year is at best
irrelevant.

You may object that it is informative. Ah yes, *informative*. For whom?
Primarily for bean counters who care about 'deliverables', 'outputs' and
other countable things. And of course for managers who care about
'productivity' and 'volume'.

These cursed little histograms invite inferences about productivity,
gaps, and publication volume that are guaranteed to be reductive and
bias-ridden. One could make the case they are actively harmful, feeding
into exactly the wrong kind of feedback loops. So why do unis do it?

## Marketed to the managerial class {#marketed-to-the-managerial-class .wp-block-heading}

Systems like Elsevier Pure are marketed to Research Managers, and every
bit of their design shows that. Only a managerial class bent on counting
output and measuring productivity is sensitive to marketing speak about
"fact-based decision making" and "unlocking your full research
potential". Universities and institutions who use Pure's
"industry-proven data model" to create automated profiles for their
researchers are making a big mistake.

Above I wrote how these public-facing histograms invite inferences that
may be harmful. Of course that's pretty much what Pure has been designed
to do. Behind the scenes, there's a plethora of ways to track metrics,
targets, and progress right down to individual researchers. Just take
the below screenshot from Elsevier's marketing materials: Pure will help
you "achieve performance goals by defining targets and tracking research
progress over time". This is the definition of corporate surveillance,
and Elsevier is of course happy to bring it to capitalist neoliberal
universities --- and in the process insinuate itself into the veins of
the system.

<figure
class="wp-block-gallery has-nested-images columns-default is-cropped wp-block-gallery-4 is-layout-flex wp-block-gallery-is-layout-flex">
<figure class="wp-block-image size-large">
<a href="https://ideophone.org/files/image-23.png"><img
src="https://ideophone.org/files/image-23-1024x382.png"
class="wp-image-5619" loading="lazy" decoding="async" data-id="5619"
srcset="https://ideophone.org/files/image-23-1024x382.png 1024w, https://ideophone.org/files/image-23-300x112.png 300w, https://ideophone.org/files/image-23-768x287.png 768w, https://ideophone.org/files/image-23-500x187.png 500w, https://ideophone.org/files/image-23.png 1277w"
sizes="(max-width: 1024px) 100vw, 1024px" width="1024"
height="382" /></a>
</figure>
</figure>

Incidentally, one has to admire the efficiency of this screenshot,
showing that three researchers whose output was "insufficient" are
"former staff", i.e. have been let go. Once again, this is straight from
the MARKETING MATERIALS of Pure, in case you were wondering. They really
leave no doubt about their goals of limitless corporate surveillance and
productivity policing.

<figure class="wp-block-image size-large">
<a href="https://ideophone.org/files/image-24.png"><img
src="https://ideophone.org/files/image-24-1024x688.png"
class="wp-image-5620" loading="lazy" decoding="async"
srcset="https://ideophone.org/files/image-24-1024x688.png 1024w, https://ideophone.org/files/image-24-300x202.png 300w, https://ideophone.org/files/image-24-768x516.png 768w, https://ideophone.org/files/image-24-447x300.png 447w, https://ideophone.org/files/image-24.png 1200w"
sizes="(max-width: 1024px) 100vw, 1024px" width="1024"
height="688" /></a>
</figure>

## Cursed little histograms {#cursed-little-histograms .wp-block-heading}

But I digress; my main beef is with the public profiles, which
thoughtlessly include these little plots wherever possible --- even in
search results! Fortunately, my own institution Radboud University
doesn't use Pure (unless the recent VSNU deal with Elsevier forces it
down our throat).

Elevating this useless histogram to such a prominent place on every
researcher's profile is the web design equivalent of "nerdview": an
ill-thought-out choice that makes very little sense to end users and is
telling of your own biases. Fellow academics are probably *the* key
audience for institutional homepages. When we look up someone's page we
do it to find a specific paper, read what they're working on, perhaps
check out recent work. We don't want to see this cursed little
histogram.

## Updates {#updates .wp-block-heading}

This post [originated as an Oct 2020 thread on
Twitter](https://twitter.com/DingemanseMark/status/1320728091279372288){target="_blank"
rel="noreferrer noopener"}, and roused quite some attention. Some
research data managers woke up from their twitter slumber to reluctantly
agree with the notion that the prominent display of this visual on every
single researchers' profile may not send the right message in a time of
*Recognition and Rewards*.

Indeed, two universities (VU Amsterdam and TU Delft) admitted outright
that the output graphs violate the principles of SF-DORA and the
Recognition and Rewards position paper, with the Free University
Amsterdam taking the lead in removing the output-per-year graphs from
their Research Portal. (Not a coincidence that the change was
spearheaded by Sander Bosch and Lena Karvovskaya, who care about open
scholarship rather than senseless stats.) Switching of this view is
incredibly simple, by the way: it is literally a matter of flicking a
single button in the Pure settings for personal profiles.

<figure class="wp-block-image size-large">
<a href="https://ideophone.org/files/EngutdsW8AIKbOq.jpg"><img
src="https://ideophone.org/files/EngutdsW8AIKbOq-1024x414.jpg"
class="wp-image-7497" loading="lazy" decoding="async"
srcset="https://ideophone.org/files/EngutdsW8AIKbOq-1024x414.jpg 1024w, https://ideophone.org/files/EngutdsW8AIKbOq-300x121.jpg 300w, https://ideophone.org/files/EngutdsW8AIKbOq-768x311.jpg 768w, https://ideophone.org/files/EngutdsW8AIKbOq-1536x622.jpg 1536w, https://ideophone.org/files/EngutdsW8AIKbOq-1600x647.jpg 1600w, https://ideophone.org/files/EngutdsW8AIKbOq.jpg 1898w"
sizes="(max-width: 1024px) 100vw, 1024px" width="1024"
height="414" /></a>
<figcaption>2021 screenshot of Pure settings showing the way to remove
the cursed little research output graph</figcaption>
</figure>

Quite a few universities didn't bother. Wageningen University signed
SF-DORA and was, last I checked, still "looking for a better assessment
system" but until that time choses to reduce its researchers to little
output graphs. Rector Frank Baaijens of TU Eindhoven coauthored the
Recognition and Rewards position paper, but his own university still
clings to the numbers game. TU Delft claims they want to recognise "a
wide spectrum of products" but their public research profiles invite
only mindless bean counting. For shame, especially since all of them
have signed SF-DORA. Actions speak louder than words.

::: {.wp-block-image}
<figure class="alignright size-full is-resized">
<a href="https://ideophone.org/files/image-51.png"><img
src="https://ideophone.org/files/image-51.png" class="wp-image-5772"
loading="lazy" decoding="async"
srcset="https://ideophone.org/files/image-51.png 929w, https://ideophone.org/files/image-51-300x293.png 300w, https://ideophone.org/files/image-51-768x751.png 768w, https://ideophone.org/files/image-51-307x300.png 307w"
sizes="(max-width: 232px) 100vw, 232px" width="232" height="227" /></a>
</figure>
:::

After much postponement, writing the final report for my NWO Veni grant
(2015-2018) turned out to be an unexpected pleasure. It made me realise
a couple of things --- key among them the role of serendipity in shaping
fundamental research.

The project was called "Towards a science of linguistic depiction".
Looking at the [publications that came
out](https://ideophone.org/papers){rel="noreferrer noopener"
target="_blank"} of it and at [imperfect indicators like
citations](http://is.gd/dingemanse_veni){rel="noreferrer noopener"
target="_blank"}, it's made some useful contributions.

So did everything go according to plan? Not at all. I had to cancel
fieldwork because of little ones, had to abandon some lofty plans for
knowledge utilisation, and had papers variously rejected for being
'vague', 'without content' or worse.

For instance, the paper I'm most proud of was rejected twice, including
by Language, where it languished 6 months in review only to get reviewer
2 call it 'charming' before slamming it (if you're an ECR, it's probably
best to avoid Language: not worth the risky wait). That charming,
twice-rejected paper [ultimately
appeared](http://doi.org/10.1017/S002222671600030X){rel="noreferrer noopener"
target="_blank"} in Journal of Linguistics, where it seems to be doing
pretty well, with a place in the top 10 most cited JL papers of the last
5 years and in the top 10 most downloaded in that journal in 2018.

## Three happy coincidences {#three-happy-coincidences .wp-block-heading}

Back to the project. While we have better answers now to the research
questions posed, and even the 'utilisation' goals were reached, how I
got there was very different from what was originally planned, with a
major role for serendipity. Three happy coincidences stand out.

First, right at the start of the project (following multiple failed
attempts) me and Tessa van Leeuwen got an NWO grant to do a "Groot
Nationaal Onderzoek" w/ thousands of participants: a wonderful
opportunity that utterly derailed all other outreach plans I'd made.
We're now working on several papers from that project; if you read
Dutch, check out this piece we wrote for Onze Taal on language,
cross-modal perception and synaesthesia: "[Taal als samenspel van de
zintuigen](https://pure.mpg.de/rest/items/item_2237451_2/component/file_2237450/content){rel="noreferrer noopener"
target="_blank"}" (PDF).

Second, at MPI, I had the opportunity to supervise a cool PhD project by
UCL graduate Gwilym Lockwood that perfectly matched the Veni project.
This is privilege kicking in: outside the Max Planck system a fully
funded PhD like this would have been unlikely.
[](https://twitter.com/DingemanseMark/status/1076084723301134336/analytics)[](https://twitter.com/DingemanseMark/status/1076084723301134336/analytics)It's
also luck: Gwilym turned out to be a productive, self-propelled student
who knew what he wanted and found a cool data science job right after
his PhD. Check out his 2017 thesis ---Talking Sense---
[here](https://figshare.com/articles/thesis/Talking_Sense_The_behavioural_and_neural_correlates_of_sound_symbolism/4872866){rel="noreferrer noopener"
target="_blank"}.

Third, a pleasant and productive synergy developed with other Nijmegen
VI-laureates Gerardo Ortega (Veni) and Asli Özyürek (Vici): the
Iconicity Focus Group, a vibrant iconicity-focused research community in
Nijmegen with lots of inspiring lectures and meetings. This culminated
in an exciting co-sponsored international workshop on 'Types of
iconicity in language use, development and processing'
[\#IFG17](https://twitter.com/hashtag/IFG17?src=hashtag_click), funded
by our three NWO grants and supported by MPI, Donders and the Centre for
Language Studies at Radboud University.

All of these things massively enhanced the project and made it possible
to meet goals set in the distant past of clueless yet hopeful
grant-writing. Add cool student projects, visiting scholars, and a
concurrent 'iconicity boom' and it's clear that success can't really be
planned.

<figure
class="wp-block-gallery has-nested-images columns-4 is-cropped wp-block-gallery-5 is-layout-flex wp-block-gallery-is-layout-flex">
<figure class="wp-block-image size-large">
<a href="https://ideophone.org/files/image-48.png"><img
src="https://ideophone.org/files/image-48-1024x576.png"
class="wp-image-5767" loading="lazy" decoding="async" data-id="5767"
srcset="https://ideophone.org/files/image-48-1024x576.png 1024w, https://ideophone.org/files/image-48-300x169.png 300w, https://ideophone.org/files/image-48-768x432.png 768w, https://ideophone.org/files/image-48-1536x864.png 1536w, https://ideophone.org/files/image-48-500x281.png 500w, https://ideophone.org/files/image-48.png 1920w"
sizes="(max-width: 1024px) 100vw, 1024px" width="1024"
height="576" /></a>
</figure>
<figure class="wp-block-image size-large">
<a href="https://ideophone.org/files/image-49.png"><img
src="https://ideophone.org/files/image-49.png" class="wp-image-5768"
loading="lazy" decoding="async" data-id="5768"
srcset="https://ideophone.org/files/image-49.png 663w, https://ideophone.org/files/image-49-213x300.png 213w"
sizes="(max-width: 663px) 100vw, 663px" width="663" height="933" /></a>
</figure>
<figure class="wp-block-image size-large">
<a href="https://ideophone.org/files/image-50.png"><img
src="https://ideophone.org/files/image-50.png" class="wp-image-5769"
loading="lazy" decoding="async" data-id="5769"
srcset="https://ideophone.org/files/image-50.png 820w, https://ideophone.org/files/image-50-300x206.png 300w, https://ideophone.org/files/image-50-768x526.png 768w, https://ideophone.org/files/image-50-438x300.png 438w"
sizes="(max-width: 820px) 100vw, 820px" width="820" height="562" /></a>
</figure>
<figure class="wp-block-image size-large">
<a href="https://ideophone.org/files/image-47.png"><img
src="https://ideophone.org/files/image-47.png" class="wp-image-5765"
loading="lazy" decoding="async" data-id="5765"
srcset="https://ideophone.org/files/image-47.png 960w, https://ideophone.org/files/image-47-300x225.png 300w, https://ideophone.org/files/image-47-768x576.png 768w, https://ideophone.org/files/image-47-400x300.png 400w"
sizes="(max-width: 960px) 100vw, 960px" width="960" height="720" /></a>
</figure>
<figcaption>Groot Nationaal Onderzoek, Gwilym Lockwood’s thesis,
Iconicity Focus Group, and ‘iconicity boom’</figcaption>
</figure>

## Serendipity {#serendipity .wp-block-heading}

Looking back at the project as a whole, it's nice to be able to point to
a coherent set of publications, and gratifying to see the growing impact
of work on ideophones & depiction. But in the report I just submitted to
NWO, I made sure to also acknowledge serendipity. Here's the final
paragraph of my report:

> I am extremely grateful to NWO for the opportunity and the freedom to
> carry out basic research using a Veni grant. As I have noted at
> several points above, while the goals have been met, the project has
> also seen considerable changes and has benefited from unexpected
> synergies. My overall conclusion is that serendipity plays an
> important role in shaping fundamental research and its applications.

The process reminded me of what the Hitchhiker's Guide to the Galaxy
says about the art of flying: "The knack lies in learning how to throw
yourself at the ground and miss". For that, you need to be "lucky enough
to have your attention momentarily distracted at the crucial moment".
Certainly some of the most fun and serendipitous collaborations have
come from me being distracted by too many new ideas at once --- often
when I was about to hit the ground (or a deadline). Make of that what
you will.

Okay, time to wrap up. Lessons learned:

1. Things will go different than planned, and that's okay (be curious).
2. Reviewer 2 doesn't get it, and that's okay (keep trying).
3. It's more fun when you do it together (collaborate).

And that's all folks. Enjoy the holidays. Take some time off, you
deserve it. So long and thanks for all the funding!