Everyone knows that linguistic forms often (seem to) have multiple meanings: Sentences may be ambiguous, words often exhibit polysemy, they may be homonymous, and grammatical markers can have a range of uses. But more recently, two additional terms have become widespread in comparative approaches: colexification (the possibility of using a word with different senses) and coexpression (the possibility, more generally, of using a linguistic form with different senses or functions).

For example, German Blatt colexifies ‘leaf’ and ‘sheet’ (e.g. ein Blatt Papier ‘a sheet of paper), and English put colexifies the meanings ‘setzen’, ‘stellen’ and ‘legen’ (taking a German perspective).

Colexification was introduced by François (2008) and further popularized by the CLICS database of worldwide colexifications (Tjuka et al. 2026: https://clics.clld.org/), while the more general term coexpression was first used by Hartmann et al. (2014). Both terms have become widely known in comparative linguistics circles, and colexification has even made it into psychology (e.g. Jackson et al. 2019) and computational linguistics (e.g. Liu et al. 2023). It appears that lexical coexpression patterns were not compared widely across languages until the early 21st century, but grammatical “polysemy patterns” were of interest to linguists earlier, discussed under the heading of “semantic maps” since the 1980s (cf. Anderson 1982; Haspelmath 2003). A very early example of a “lexical semantic map” that got famous is Hjelmslev’s comparison of Danish træ ‘tree, wood’ with other languages (cited and updated by many authors, e.g. François 2022); and an early example of a semantic map in typology is from my 1987 paper on anticausatives and related voice functions:

In addition to colexification (for the coexpression of lexical meanings), one may want to have a parallel term, cogrammification, for the coexpression of grammatical meanings, as in the anticausative-related semantic map above (I proposed this term in Haspelmath 2023: §1, also encompassing what is often described as syncretism).

Now two questions arise: (1) What is the nature of the meanings or senses or functions that are coexpressed (or disexpressed, i.e. distinguished, as the case may be) in languages; and (2) what is the nature of the forms that may occur with multiple meanings?

Ad (1): When a word has multiple meanings or senses, it is typically said to be polysemous, and coexpression patterns used to be called “polysemy patterns”. However, “polysemy” is not what is at stake in most cases of colexification. Danish træ ‘tree, wood’ is not (obviously) polysemous in the sense of having multiple meanings: Wood is made of trees, and it is normal in languages that material and its source are expressed by the same word (e.g. wheat as a plant or as the ingredient for bread, or wool as a part of a sheep’s skin or as the ingredient for clothes). Similarly, English wood is not polysemous just because Spanish and Mwotlap have two different words corresponding to it (depending on its use for fire or for construction). Comparative linguists have always known this, and it is for this reason that the terms colexification and coexpression caught on quickly. Polysemy is the existence of two meanings within one language (it is a notion of p-linguistics), but colexification is broader: It also encompasses cases where there is just a single meaning within a language, but from the perspective of another language, one may distinguish different senses. For example, the English verb put (mentioned earlier) is not polysemous, but from the German perspective (where the three postures setzen, stellen and legen are distinguished), one can say that put colexifies these three meanings.

The meanings that are colexified (or cogrammified, or more generally coexpressed) are therefore not language-particular meanings, but comparative concepts: They are comparison meanings. François (2022) uses the term sense for such meanings, and he makes it very clear that they are established by comparison of languages, not language-internally:

In another passage (2022: 90), he refers to them as “potential meanings”, forming an “etic grid”:

“When comparing how languages cut up the semantic space, one efficient approach consists in breaking up a given lexical domain into a number of potential meanings, independent of their expression in individual languages. The “etic grid” formed by these potential senses then serves as a background against which emic categories of particular languages can be compared.”

The distinction between “etic categories” and “emic categories” is the same as the distinction between comparative concepts and language-particular descriptive categories (see Haspelmath 2010: 678, 681). Among typologists, the distinction between comparative and descriptive notions has been generally accepted, but the term “comparative concept” is not widely applied to semantic typology yet (for example, it is missing in Koptjevskaja-Tamm et al. 2016 and Brochhagen & Boleda 2022). However, more recently, it has increasingly been recognized that the comparison meanings of lexification typology are comparative concepts in the same way as the grammatical concepts that I highlighted in the 2010 article. For example, Snee et al. (2026) introduce the notion of a concept list as follows:

“a concept list typically defines a set of language-independent comparative concepts, which serve as standardized reference points for cross-linguistic comparison (Haspelmath 2010). For each comparative concept, researchers elicit corresponding word forms in the target languages. Although one word form is usually elicited for a comparative concept in a list, in practice more than one may be reported per meaning…”

For grammatical coexpression patterns, linguists have also used term such as “functions” (Haspelmath 1997) or “uses” (Bybee et al. 1994), but it is clear that what is meant is functional-semantic comparative concepts. Semantic analysis is typically thought of as language-internal semantic analysis, involving language-particular meanings. But comparison of meanings is different, in that it works with “potential meanings” (to use François’s expression), and I have therefore been wondering whether we should not talk about “lexification typology” (rather than “semantic typology”) and “coexpression diagrams” (rather than “semantic maps”, cf. Haspelmath 2023: §4.1).

Next, ad (2), the nature of the forms: What exactly is it that can coexpress multiple comparison meanings? For cogrammification, I have said that it is grammatical morphs: cogrammification is the mapping of multiple meanings onto a single grammatical morph (Haspelmath 2023: §2), or the “expression of either A or B in a grammatical marker” (§3.1). I use “grammatical morph” and “grammatical marker” synonymously, and a “grammatical meaning” is a meaning expressed by a grammatical marker (for the definition of the latter, see the Grammaticon entry: https://grammaticon.clld.org/concepts/1).

Analogously, I would say that colexification is the mapping of multiple meanings onto a single root (= lexical morph). The notion of a lexeme is defined in terms of the notion of a root (see the Grammaticon entry: https://grammaticon.clld.org/concepts/513).

But what exactly is a root? I discussed this term in an earlier DLC blogpost, and also in a 2025 paper: “Roots and root classes in comparative grammar”). In this paper, I say that noun-verb pairs such as English hammer (verb) and hammer (noun) are two different roots, so that one wouldn’t say that the two meanings ‘tool with handle and head for impacting’ and ‘use a hammer’ are colexified.

However, other linguists use the term “root” differently – they would say, for instance, that English hammer is a single categoryless root that gets its different meanings and syntactic properties from the constructional/syntactic environment. This would then imply that ‘impacting tool’ and ‘use a hammer’ are colexified in hammer. I have not seen anyone actually saying this, though, as the colexification literature tends to focus on noun colexification or verb colexification. The literature on “noun-verb conversion” (e.g. Bauer & Valera 2026) has not made any contact with the colexification literature, it seems.

But the question of root identification arises in many other cases as well. Take the example of English trunk, which can mean ‘upright part of a tree’, ‘elephant proboscis’, or ‘storage compartment of a car’ (cf. François 2022: 92) – these are very different from each other, so that a polysemy analysis is difficult: It seems that we need to say that there are three different homonymous English roots with the shape /tɹʌŋk/. If they are different forms, they do not colexify different meanings.

However, how do we distinguish in general between a situation of homonymous root sets and a situation of polysemous roots? This question is often discussed in semantics textbooks (e.g. Kroeger 2019: §5.3.3), but there does not seem to be any general solution. Many linguists say that polysemy and homonymy cannot be distinguished, and many others say that polysemy cannot be distinguished from indeterminacy of meaning (e.g. Geeraerts 2001).

List et al. (2026) say in a recent blogpost that the polysemy/homonymy distinction is more an explanatory distinction than a descriptive distinction:

“Polysemy is usually explained as the result of semantic change when individual forms acquire more senses, while homophony is explained as the result of sound change leading to phonological merger in individual forms of a language.” (https://calc.hypotheses.org/9547)

They thus prefer to apply the term colexification to forms regardless of whether they are single polysemous roots or multiple homonymous roots, but they note that other authors see things differently.

Notably, List et al. point out that there was a controversy at a recent workshop (ALT Conference Lyon, workshop on the dynamics of lexification) about the use of colexification: Some colleagues preferred a broader use (including homonymous roots), while others preferred a narrower use (including only indeterminate/vague and polysemous roots).

List et al. have used colexification as “a cover term for all cases where one form has multiple senses in a given language”, and where by “one form” they mean a single sequence of segments (i.e. a single shape, using the terminology of Haspelmath 2020). By contrast, in François’s (2022: 95) definition, colexification refers to the possibility of expressing two distinct senses “using the same lexical form” (also François 2008: 170). In subsequent discussions, François has made it clear that by “lexical form” he means a form with a single meaning or with several related meanings, excluding cases of homonymy in which the meanings are unrelated.

What should one do under these circumstances? We have a beautiful new term, but it seems to be ambiguous already! It seems to me that we have to live with a certain degree of indeterminacy here, because I fear that we will never agree on how to distinguish indeterminacy/polysemy/homonymy. At least for root pairs such as hammer (noun) / hammer (verb), I think we need to posit two different roots (as noted above, and as explained in my 2025 paper), and many people will want to do it also in many other cases. Fortunately, it probably does not matter so much, because cases of “true homonymy” (where the two words have different diachronic origins) are not very common, compared to cases of polysemy (including trunk, whose very different meanings all arose from a single etymon).

Finally, let me briefly mention that for cogrammification, very similar issues arise. In inflectional paradigms, cogrammification has often been called “syncretism”, and the distinction between “systematic syncretism” (similar to indeterminacy/polysemy) and “accidental syncretism” is notoriously hard (or impossible) to make (see, e.g., Caha 2019, discussed here, and Esher 2026).

References

Anderson, Lloyd B. 1982. The “perfect” as a universal and as a language-particular category. In Hopper, Paul J. (ed.), Tense-aspect: Between semantics and pragmatics, 227–264. Amsterdam: Benjamins.

Bauer, Laurie & Valera, Salvador. 2026. On the borders of conversion. Folia Linguistica. (doi:10.1515/flin-2025-0165) (https://www.degruyterbrill.com/document/doi/10.1515/flin-2025-0165/html) (Accessed August 23, 2026.)

Brochhagen, Thomas & Boleda, Gemma. 2022. When do languages use the same word for different meanings? The Goldilocks principle in colexification. Cognition 226. 105179. (doi:10.1016/j.cognition.2022.105179)

Bybee, Joan L. & Perkins, Revere & Pagliuca, William. 1994. The evolution of grammar: Tense, aspect and modality in the languages of the world. Chicago: The University of Chicago Press.

Caha, Pavel. 2019. Syncretism in morphology. Oxford Research Encyclopedia of Linguistics 2019. (doi:10.1093/acrefore/9780199384655.013.552)

Esher, Louise. 2026. Syncretism. The Wiley Blackwell Companion to Diachronic and Historical Linguistics, 1–23. Wiley. (doi:10.1002/9781119898023.wbcdl096) (https://onlinelibrary.wiley.com/doi/abs/10.1002/9781119898023.wbcdl096)

François, Alexandre. 2008. Semantic maps and the typology of colexification: Intertwining polysemous networks across languages. In Vanhove, Martine (ed.), From polysemy to semantic change: Towards a typology of lexical semantic associations (Studies in Language Companion Series 106), 163–216. Amsterdam: Benjamins.

François, Alexandre. 2022. Lexical tectonics: Mapping structural change in patterns of lexification. Zeitschrift für Sprachwissenschaft 41(1). 89–123. (doi:10.1515/zfs-2021-2041)

Geeraerts, Dirk. 2001. The definitional practice of dictionaries and the cognitive semantic conception of polysemy. Lexicographica 17. 6–21.

Hartmann, Iren & Haspelmath, Martin & Cysouw, Michael. 2014. Identifying semantic role clusters and alignment types via microrole coexpression tendencies. Studies in Language 38(3). 463–484. (doi:https://doi.org/10.1075/sl.38.3.02har)

Haspelmath, Martin. 1987. Transitivity alternations of the anticausative type (Arbeitspapiere Des Instituts Für Sprachwissenschaft N.F. Nr. 4). Cologne: Universität zu Köln. (http://doi.org/10.5281/zenodo.225196)

Haspelmath, Martin. 2003. The geometry of grammatical meaning: Semantic maps and cross-linguistic comparison. In Tomasello, Michael (ed.), The new psychology of language (Volume 2), 211–242. Mahwah, NJ: Lawrence Erlbaum. (doi:https://doi.org/10.5281/zenodo.831410)

Haspelmath, Martin. 2010. Comparative concepts and descriptive categories in crosslinguistic studies. Language 86(3). 663–687. (doi:10.1353/lan.2010.0021)

Haspelmath, Martin. 2023. Coexpression and synexpression patterns across languages: Comparative concepts and possible explanations. Frontiers in Psychology 14. (doi:10.3389/fpsyg.2023.1236853) (https://www.frontiersin.org/journals/psychology/articles/10.3389/fpsyg.2023.1236853/full)

Haspelmath, Martin. 2025. Roots and root classes in comparative grammar. In Levkovych, Nataliya & Nintemann, Julia & Vorholt, Maike (eds.), Exploring structures in languages and language contact, 307–324. Berlin: De Gruyter Mouton. (https://www.degruyterbrill.com/document/doi/10.1515/9783111496436-012/html)

Jackson, Joshua Conrad & Watts, Joseph & Henry, Teague R. & List, Johann-Mattis & Forkel, Robert & Mucha, Peter J. & Greenhill, Simon J. & Gray, Russell D. & Lindquist, Kristen A. 2019. Emotion semantics show both cultural variation and universal structure. Science 366(6472). 1517–1522. (doi:10.1126/science.aaw8160)

Koptjevskaja Tamm, Maria & Rakhilina, Ekaterina & Vanhove, Martine. 2016. The semantics of lexical typology. In Riemer, Nick (ed.), The Routledge handbook of semantics, 434–454. London: Routledge.

Kroeger, Paul R. 2019. Analyzing meaning (2nd edition). Language Science Press. Berlin: Language Science Press. (doi:10.5281/zenodo.2538330) (https://langsci-press.org/catalog/view/231/1728/1363-1)

List, Johann-Mattis & Bocklage, Katja & Georgakopoulos, Thanasis. 2026. The ambiguity of the colexification term. Computer-Assisted Language Comparison in Practice: Tutorials on Computational Approaches to the History and Diversity of Languages 9(2). 85–92. (doi:10.15475/calcip.2026.2.2)

Liu, Yihong & Ye, Haotian & Weissweiler, Leonie & Pei, Renhao & Schuetze, Hinrich. 2023. Crosslingual Transfer Learning for Low-Resource Languages Based on Multilingual Colexification Graphs. In Bouamor, Houda & Pino, Juan & Bali, Kalika (eds.), Findings of the Association for Computational Linguistics: EMNLP 2023, 8376–8401. Singapore: Association for Computational Linguistics. (doi:10.18653/v1/2023.findings-emnlp.562) (https://aclanthology.org/2023.findings-emnlp.562/) (Accessed August 23, 2026.)

Snee, David & Ciucci, Luca & List, Johann-Mattis. 2026. Variation in language phylogenies may result from variation in concept translation. Journal of Language Evolution 11. lzag005. (doi:10.1093/jole/lzag005)

Tjuka, Annika & Forkel, Robert & Rzymski, Christoph & List, Johann-Mattis. 2026. CLICS4: An Improved Database of Cross-Linguistic Colexifications. Passau: MCL Chair at the University of Passau. (doi:10.5281/zenodo.16900179) (https://clics.clld.org/)

 

 

---

The text only may be used under licence Creative Commons Attribution 4.0 International. All other elements (illustrations, imported files) are “All rights reserved”, unless otherwise stated.

---

---