Many linguists are aware that two of the most basic concepts of our field, “word” and “lexicon”, are used in confusing ways. But can we do something about this, other than pointing out the confusions once in a while? I now think we can, by defining “word” in a precise way (Haspelmath 2023), and by replacing “lexicon/lexical” in by one of four other terms if necessary (set of words, lexemicon, inventorium, mentalicon). For mnemonic reasons and because it was fun, I published my paper about “word” in the journal WORD, and my paper about “lexicon” was recently published in the journal “Lexique” (Haspelmath 2024b).

We inherited a lot of concepts and technical terms from our ancestors, and while some of them are clearly outdated (e.g. we no longer talk about “consonant letters”, but about “consonant sound segments”, or “phonemes”), others continue to be used even if we know that their use is unclear. “Word” and “lexicon” are not just vague, but they invoke multiple stereotypes, and many linguists have formulated theories that crucially rely on these concepts (e.g. “lexicalist” theories, “word-based” morphology). So what exactly these terms are taken to mean is often important, but it is also often unclear.

Mark Aronoff and Ray Jackendoff have repeatedly pointed out some of these problems, and they offered their solutions – but as these come with particular theoretical choices (“morphology by itself”, “Parallel Architecture”), linguists who were not persuaded by these choices have often ignored their terminological points. However, one need not be sympathetic to any particular view of language or grammar to realize that the term “lexicon/lexical” has multiple senses – and that this is a serious practical problem. Jackendoff (2013: 74) lists three different senses of “lexical item”:

However, he presents the issue as if “lexicon” and “lexical item” were necessary concepts given in advance and that the question is how we characterize or identify them. Similarly, in his (2002b) paper, he asks in the title: “What’s in the lexicon?” But “lexicon” is not a necessary term in linguistics (it did not exist before Bloomfield 1933), so it’s odd to start out from it rather than from the most basic concepts.

When a technical-sounding term is used in multiple senses, one way to deal with the situation is to reject the term as unsuitable. I did this with the term markedness in my 2006 paper, and I was partly successful. That paper is still widely cited, and when a linguist uses the term, they often add a footnote apologizing for it (with reference to Haspelmath 2006) and (crucially!) explaining in what sense they use the term.

For some time, I wanted to treat terms like “inflection”, “derivation”, “lexeme”, “lexicon”, and “word” in the same way (rejecting them as unsuitable), but I now think that as these terms will continue to be used anyway, it is better to propose precise definitions for them (for “inflection/derivation”, see Haspelmath 2024a). So I now define ‘word’ in an unnatural way, but the definition is quite precise.

For “lexicon” and “lexical item”, I think that the best approach is to come up with four different terms for the different senses, as follows (for details, see the new 2024b paper)

(i) lexical entity as word-form (or simply word)
(ii) lexical item as lexeme (an abstract element based on a root)
(iii) lexical item as inventorial item (an item of the inventorium)
(iv) lexical item as mental item (an item of the “mental lexicon” or mentalicon)

Stereotypically, a “word” is either (i) form that occurs between two spaces in writing, or (ii) a form that is listed in a dictionary, or (iii) a form that must be remembered or stored when learning a foreign language. However, as Jackendoff (and many others) have observed, these notions need not coincide, and there is a fourth notion: (iv) a form that is actually remembered by a speaker.

For the first notion, we can also say word-form; for the second, we have been using the term lexeme for some time (especially since Matthews 1972; 1974); for the third, some authors have said “listeme”; and for the fourth, one can say “mental item”. Jackendoff generally uses the term “lexical item” for forms that ARE listed in long-term memory (2c above), but conceptually, these are distinct from the idiosyncratic forms that MUST BE listed (2b above, what I call inventorial items, or inventoremes). In my view (and here I differ from Jackendoff), linguists are mostly concerned with the inventorium, not “the mental lexicon”, because the latter is different for different speakers. Linguistics is primarily about studying social conventions, not about mental grammars.

The key observation is that the misunderstandings in linguistics seem to have their roots in our traditional “word” stereotype: Since we most often encounter tangible words in writing (between spaces in our Western languages), and also often in dictionaries (alphabetically arranged until recently), we treat both as if they were somehow the same thing. But of course, many or most words that occur in written texts are not found in a dictionary (at least in texts from languages like Latin or Turkish), and many forms that must be remembered are not words (because we must learn a huge number of idioms, collocations and clichés).

Thus, I think of the terms “lexical item” and “lexicon” as similar to the older term “consonant letter” (referring to a consonant sound segment). In many cases, these terms do not do much damage, but in a technical scientific context, it is better to replace them by more precise terms.

The four different concepts presented here are completely uncontroversial, I think. Linguists may differ in which aspects of languages they focus on, and which concepts they rely on for their theories. But at a basic level, it is clear that we want to have (at least) the concepts “word”, “lexeme”, “inventorium”, and “mentalicon”. Thus, I am not advancing any particular theory, but merely pointing out the need for clearer terminology. It is a modest contribution.

References

Bloomfield, Leonard. 1933. Language. New York: H. Holt and Company.

Haspelmath, Martin. 2006. Against markedness (and what to replace it with). Journal of Linguistics 42(1). 25–70. (doi:10.1017/S0022226705003683)

Haspelmath, Martin. 2023. Defining the word. WORD 69(3). 283–297. (doi:10.1080/00437956.2023.2237272)

Haspelmath, Martin. 2024a. Inflection and derivation as traditional comparative concepts. Linguistics 62(1). 43–77. (doi:10.1515/ling-2022-0086)

Haspelmath, Martin. 2024b. Four kinds of lexical items: Words, lexemes, inventorial items, and mental items. Lexique: Revue en Sciences du Langage (34). 71–95. (doi:10.54563/lexique.1737)

Jackendoff, Ray. 2002a. Foundations of language: Brain, meaning, grammar, evolution. Oxford: Oxford University Press.

Jackendoff, Ray. 2002b. What’s in the lexicon? In Nooteboom, Sieb & Weerman, Fred & Wijnen, Frank (eds.), Storage and Computation in the Language Faculty (Studies in Theoretical Psycholinguistics), 23–58. Dordrecht: Springer Netherlands. (doi:10.1007/978-94-010-0355-1_2)

Jackendoff, Ray. 2013. Constructions in the Parallel Architecture. In Hoffmann, Thomas & Trousdale, Graeme (eds.), The Oxford handbook of Construction Grammar, 70–92. Oxford: Oxford University Press. (https://doi.org/10.1093/oxfordhb/9780195396683.013.0005)

Matthews, Peter H. 1972. Inflectional morphology. Cambridge: Cambridge University Press.

Matthews, Peter H. 1974. Morphology. Cambridge: Cambridge University Press.

---

---