Additional background + goals

This workshop is about words, their representation, organization, and usage by humans and machines. A good lexical component is vital for any natural language system. There is hardly any task (mental action) one can perform without these building blocks (analysis/synthesis, paraphrasing, or translation). Words are very powerful, yet, amazingly versatile. While a single word may capture in a nutshell a world of knowledge, its combination with other words (concepts) may support communication, memorization(storage/retrieval), and even thinking, an activity that consists largely of linking ideas (propositions), and, in combining concepts (standing for entities of various sorts: objects, events, processes) for both of which we have created words. 

Words can entertain, reveal, or hide. Sharper than razor blades, they can make us also laugh or cry. No doubt, words are important. They are everywhere, multifarious, and somehow mysterious —after all, what is a word?— which is probably why so many scholars have been attracted by them. And yet, despite all the efforts, there are still so many questions we have not asked, let alone found the right answer for. For example, 

• How are words represented in the human mind?

• How are they organized, accessed, and remembered, i.e., learned?

• What does the mental lexicon look like? (For a curated list of work devoted to this notion, see: http://pageperso.lif.univ-mrs.fr/~michael.zock/Docs/MZ_Refs_ML.pdf ).

• What is its topology, and how can we build such a map?

• How do words convey subliminal messages, and how can we reveal these hidden meanings?

• How can we boost information to make sure that the needed words appear in the upper part of the list of elements among which an author may have to choose?

• Are neural-based word embeddings well suited to do this job, and are they better than, say, co-occurrence extractors?

Words in books and our mind (human brain) are not the same: while in books/dictionaries, they exist as holistic entities, organized alphabetically (or, by radicals and number of strokes, as in the case of Chinese), in the human brain they are decomposed, the mental lexicon being a high-dimensional network. Words have been studied by many researchers, but, of course, not always with the same goal. While psychologists have focused on the process (time course leading from meaning to form), linguists or lexicologists have studied the product, trying to build a map within which the search takes place. This has led to a variety of resources (wordnets, thesauri, knowledge graphs, rime dictionaries) reflecting different knowledge states. While these two approaches are fundamentally different, they are complementary. Hence, they should be reconciled and integrated into a single lexical framework (Stella, 2018). 

One of the goals of this workshop is to achieve precisely that. Another goal is to gather people with different backgrounds to learn from each other, be it to improve their work or to build a resource taking into account the complementarity of the various perspectives. There is a need for this, and one may wonder why there has been so little interaction between the different communities: network scientists, psychologists, neuroscientists, linguists/lexicographers ...? Obviously, they all do have many points and interests in common. For example, nowadays, we speak about wordnets, hence, the link between neuroscience (Sporn, 2010; Bullwinkle & Sporn, 2009; Spitzer, 1999) and the graph community seems obvious (Siew et al. 2019; Vitevitch et al.,2014; Vitevitch, 2008). Likewise, one could take inspiration from this kind of work to explore the very notion of the mental lexicon: what does the map look like (Aitchison, 2003; Bonin, 2004; Papafragou et al., 2022; Zock et al. 2021) and how to build it (Zock, 2022). 

As we live in a fast-moving world, it is hard for anyone to stay on top of the wave. Yet there are always new ideas, and quite a few of them remain hidden and ignored. For example, who is aware of Stella's notion of the multiplex structure of the mental lexicon (Stella et al. 2018), the work on meaning, or, more precisely, the way linguistic forms 'acquire' it (Bolognesi, 2020). There is also a proposal of 'doing' cognitive lexicography (Ostermann, 2015). And, last, but not least, there is ChatGPT, which has generated a lot of expectations (de Schryver, G.-M. & D. Joffe. (2023). All these ideas are important and deserve to be better known. 

It is intriguing to see that we are still unable to explain why or how people (generally) manage to access words on the fly, i.e., in no time. Likewise, it is interesting to see how differently psychologists and lexicographers view the very notion of words, their organization, and access. 

Psychologists address the issue in terms of activation, i.e., the route leading from meaning to sound (Levelt et al., 1999), while computational lexicographers (Fellbaum, 1998) address it in terms of navigation. Having defined the ‘mental lexicon’ in terms of associations (Debrenne, 2010; de Deyne et coll. 2016; Lafourcade, 2015; Meara, 2009), they offer a partial map allowing users to find the word via navigation. The two communities work on different planes (vertical/horizontal) and on different time scales. 

Psychologists (Kerr, et al., 2023) describe the way words are synthesized in real-time, i.e., online processing, while computational linguists (Hogan et al., 2021) present a partial map of the ‘mental lexicon’, allowing for offline processing (navigation). Yet, given their respective approaches neither nor can provide us the map, i.e., the collective mental lexicon, containing at the various levels (meaning, sound, form), the needed information, so that a user can find the word he is looking for, regardless of her knowledge state. The reasons for this failure are due to granularity, scope, and the ever-changing knowledge states. Psychologists work on an extremely small scale (generally less than 100 words), with minute details, but within a connectionist framework. Alas, being non-symbolic, these networks cannot be interpreted by a human (ordinary dictionary) user. 

Lexicographers work at a coarser grain level. They try to build lexical networks akin to WordNet. Alas, so far, no one has been able to build for a given community (language) a complete map of their (collective) mental lexicon. Also, linguists do not consider the user's meta-knowledge and the dynamically changing knowledge states (knowledge at the onset of the search). This is something psychologists are sensitive to. As one can see, the concerns and results of these two communities are complementary. Hence, they could benefit from each other, but actually, we need more than that. There is the problem of knowledge gaps (we often ignore the existing), the problem of awareness (we do not know what is needed), and the problem of strategies (goal: what is the user looking for? Why does she change her strategy?). To make progress, we need to be aware of all of this, and to this end, we must meet and talk to each other. This is something that is still too often lacking.

References

Aitchison, J. (2003). Words in the Mind: an Introduction to the Mental Lexicon. Oxford: Blackwell.

Bolognesi, M. (2020). Where words get their meaning, Amsterdam, John Benjamins Publishing

Bonin, P. (2004). Mental Lexicon: Some Words to Talk About Words. Nova Science Publishers

Castro, N. & Stella, M. (2018). The multiplex structure of the mental lexicon influences picture naming in people with aphasia. Journal of Complex Networks, 7(6), 913-931.

Debrenne, M. (2010). Dictionary of word associations for French. Volume 1: From stimulus to reaction; Volume 2: From reaction to stimulus. (In Russian and French). Novosibirsk 

de Deyne, S., Verheyen, S. & Storms, G. (2016). Structure and organization of the mental lexicon: A network approach derived from syntactic dependency relations and word associations. In Mehler, A., Lücking, A., Banisch, S., Blanchard, P. & Job, B. (Eds.). Towards a theoretical framework for analyzing complex linguistic networks (pp. 47–79). Berlin: Springer.

de Schryver, G.-M. & D. Joffe. (2023). ‘The End of Lexicography, Welcome to the Machine: On How ChatGPT Can Already Take over All of the Dictionary Maker’s Tasks.’ Paper presented at the 20th CODH Seminar. Center for Open Data in the Humanities, Research Organization of Information and Systems, National Institute of Informatics, Tokyo, Japan, 27 February 2023. https:// youtu.be/watch?v=mEorw0yefAs 

Fellbaum, C. (Ed.), (1998). WordNet: An electronic lexical database and some of its applications. Cambridge: MIT Press. 

Hogan, A., Blomqvist, E., Cochez, M., de Melo, G., Gutierrez, C., Kirrane, S., Labra Gayo, J. E., Navigli, R., Neumaier, S., Ngonga Ngomo, A-C., Polleres, A., Rashid, S. M., Rula, A., Schmelzeisen, L., Sequeda, J., Staab, S., Zimmermann, A., & d Amato, C. (2021). Knowledge Graphs. ACM Computing Surveys, 54, 1 - 37. https://doi.org/10.1145/3447772

Kerr, E., Ivanova, B., & Strijkers, K. (2023). Lexical access in speech production. In Hartsuiker, R. J., & Strijkers, K. (Eds.). Language Production. Taylor & Francis. pp. 32-65.

Lafourcade, M. & Joubert, A. (2015). TOTAKI: A help for lexical access on the TOT Problem. In Gala, N., Rapp, R. & Bel-Enguix, G. (Eds). Language Production, Cognition, and the Lexicon. Festschrift in honor of Michael Zock. Dordrecht, Springer, pp. 95-112

Levelt W., Roelofs A. & Meyer, A. (1999). A theory of lexical access in speech production. Behavioral and Brain Sciences, 22, 1-75. 

Meara, P. (2009). Connected Words: word associations and second language lexical acquisition. Amsterdam: John Benjamins.

Ostermann C. (2015). Cognitive Lexicography: A New Approach to Lexicography Making Use of Cognitive Semantics. De Gruyter

Papafragou, A., Trueswell, J. C., & Gleitman, L. R. (Eds.). (2022). The Oxford Handbook of the Mental Lexicon. Oxford University Press.

Siew, C., Wulff, D., Beckage, N., & Kenett, Y. N. (2019). Cognitive Network Science: A review of research on cognition through the lens of network representations, processes, and dynamics. Complexity, 2019

Stella M., Beckage N., Brede M. & De Domenico M. (2018). Multiplex model of mental lexicon reveals explosive learning in humans. Scientific reports. 8(1):2259. pmid:29396497

Vitevitch, M., Goldstein, R., Siew, C. & N. Castro (2014). Using complex networks to understand the mental lexicon. Yearbook of the Poznan Linguistic Meeting, pp. 119–138 

Zock, M., de Deyne, S., Stella, M. & Pirrelli, V. (2021). The mental lexicon, blueprint of tomorrow's electronic dictionaries: cognitive aspects of the lexicon. Frontiers, Special Issue in AI and Psychology, https://www.frontiersin.org/research-topics/17890/the-mental-lexicon-blueprint-of-the-dictionaries-of-tomorrow-cognitive-aspects-of-the-lexicon

Zock, M. (2023). The mental lexicon: A blueprint for the dictionaries of tomorrow? Frontiers in Artificial Intelligence, 5, 1-4. DOI 10.3389/frai.2022.1027392.