Willis (2007) documented three historical changes in which a grammatical item was reanalyzed into a full lexical item — Bulgarian nešto (indefinite pronoun 'something' → noun 'thing'), Welsh eiddo (pronominal possessive → noun 'property'), and Welsh nôl (complex preposition yn ol 'after' → verb 'fetch') — and proposed to recognize in them a distinct type of degrammaticalization, 'syntactic lexicalization', produced by the same mechanisms as grammaticalization (structural reanalysis and pragmatic inferencing) but in the reverse direction. He explained the apparent cross-linguistic rarity of the type by two constraints: the source form must have become grammatically or semantically isolated from its paradigm, and its surface shape must be morphophonologically plausible as a form of the target lexical category. The open problem left by the study is whether this type and its two constraints hold beyond the three founding cases: documenting further instances across languages (or establishing their absence), testing both constraints on a broader sample, adjudicating contested candidates — especially modal-to-lexical shifts, where putative degrammaticalization may instead be survival of pre-grammatical lexical uses — and fixing the enabling roles of contact, analogy, dialect variation, exaptation, and divergence. Later literature (2009–2026) has produced typological reframings (Norde's degrammation typology; constructionalization), rival generalizations (Viti 2015), and a steady stream of new individual cases (Dutch fake morphemes, Polish prze-, Italian presso/pressi, Abruzzese impersonal pronouns, Spanish enviar a + infinitive), but no systematic cross-linguistic test of the two-constraint hypothesis and no resolution of the modal candidates; the core remains open.

Background

Grammaticalization — the historical process by which lexical items acquire grammatical functions, as when English 'going to' developed into a future marker — is widely held to be largely unidirectional (lexical → grammatical). Whether the reverse direction, degrammaticalization, genuinely occurs was one of the most contested issues of the grammaticalization debates around 2000: critics challenged the theoretical framework itself (Campbell 2000; Newmeyer 2000), and Haspelmath (1999) argued that degrammaticalization is 'by and large impossible' because speakers innovate toward extravagant explicitness, with no counteracting 'maxim of anti-extravagance'. Within this debate, Willis (2007) analyzed three changes in which a grammatical item became a full lexical item of an inflected category:

  1. Bulgarian nešto: the indefinite pronoun něčǐto 'something' (Old Church Slavonic lineage) developed, alongside a pronominal use, into a common noun 'thing' with complete nominal morphology (plural nešta, definite neštoto/neštata), after the loss of the interrogative čito and of case morphology removed the paradigmatic cues tying the form to the pronominal system.
  2. Welsh eiddo: a stressed pronominal possessive (from Common Celtic */esjo/ 'his') became the noun 'property', which now accepts adjectives, demonstratives, and quantifiers — modification impossible for ordinary pronouns.
  3. Welsh nôl: the Middle Welsh complex preposition yn ol 'after' (yn 'in' + the noun ol 'track') was structurally reanalyzed as the verbnoun (the uninflected citation form of Welsh verbs) nôl 'fetch', through bridging contexts in which 'went after X' contextually implied 'went to fetch X'; inflected verb forms are attested from the sixteenth–seventeenth centuries, with dialectal outcomes northern nôl versus southern hôl.

Willis argued these constitute a distinct type, syntactic lexicalization: the changes are driven by the same mechanisms as grammaticalization — structural/category reanalysis (new parse of an unchanged surface string) and pragmatic inferencing (hearers' contextual implications becoming conventional meaning) — display continuity between the old and new analyses (surface strings ambiguous between the two parses), and do not rely on derivational morphology, which distinguishes them from ordinary morphological lexicalization (conversion or zero-derivation, where a word is simply used in another word class). To explain why such changes are cross-linguistically rare, he proposed two constraints: (i) isolation — the source form must have lost its paradigmatic or semantic ties to its grammatical function (eroded paradigms, lost cognate forms); (ii) morphophonological plausibility — its surface shape must fit a possible word-form of the target lexical category, so that learners can converge on a lexical analysis (nešto has the shape of a neuter noun; nôl has the shape of a zero-marked verbnoun). The proposal connects to exaptation — Lass's (1990) term for the recruitment of marginal linguistic 'junk' for new functions — and to the idea that category reassignment across discontinuous intergenerational transmission is what allows learners to reanalyze at all (Janda 2000). A parallel authenticity problem concerns modal verbs: Beths (1999) claimed that the English modal dare was reanalyzed as a full lexical verb, a putative degrammaticalization whose status remains contested because dare's full-verb uses may simply be survivals of its pre-grammatical lexical stage rather than innovations. Norde's (2009) monograph subsequently established the standard typology of counter-directional change — degrammation (shift from a minor to a major word class), debonding (bound form becomes free), deinflectionalization (inflectional affix becomes derivational) — and treats Willis's Welsh case under degrammation, without testing Willis's two constraints.

Problem Statement

The open question arising from Willis (2007)'s central claims — that syntactic lexicalization constitutes a new type of degrammaticalization, that it is constrained by two licensing conditions, and that it therefore appears to be cross-linguistically rare — is to establish whether syntactic lexicalization is a genuine, recurrent type of degrammaticalization and what licenses it. Its jointly linked components, elaborating that single core objective, are:

(1) Generality: document further instances across languages — beyond Bulgarian nešto, Welsh eiddo, and Welsh nôl — in which a grammatical item has been reanalyzed into a full lexical item by structural reanalysis and pragmatic inferencing meeting Willis's diagnostics (mechanism identity with grammaticalization, continuity/structural ambiguity between the old and new parses, no reliance on derivational morphology), or establish how rare the type is and in what kinds of languages or subsystems it occurs.

(2) Constraints: test on that evidence whether the two proposed limiting conditions hold generally — must the source form have become grammatically or semantically isolated from its paradigm, and must its surface form be morphophonologically plausible for the target lexical category, for the reanalysis to go through — or do genuine cases exist that violate either constraint?

(3) Contested candidates: adjudicate the proposed modal-to-lexical shifts (e.g., English dare, per Beths 1999): are they genuine syntactic lexicalization, survivals of pre-grammatical lexical uses, or other processes such as constructional lexicalization or pragmaticalization?

(4) Enabling factors: determine the precise roles of language contact, analogy, and dialectal phonological variation in the attested reanalyses (for example, the northern nôl versus southern hôl split and possible influence from English), together with the chronological details of when pragmatic inferences became conventionalized and when paradigm decay made reanalysis available.

(5) Theoretical framing: evaluate whether exaptation-style accounts or gradual-divergence accounts (split of grammaticalized and ungrammaticalized variants) better capture the different subcases, including how learner-driven reanalysis across transmission discontinuity fits either account.

All five components serve one objective: to decide, on a documented cross-linguistic basis, whether syntactic lexicalization is a well-defined, predictably constrained type of counter-directional change.

Deciding whether the type is well-defined and predictably constrained rests on a documented cross-linguistic basis: qualifying further cases exhibited (or their genuine absence established), together with a verdict on each of the two constraints (upheld, violated, or requiring refinement, including the outright refutation of a constraint). Components (3)–(5) are further sub-questions of the same objective.

References

  1. Willis, David. 2007. Syntactic lexicalization as a new type of degrammaticalization. Linguistics 45(2): 271–310. DOI: 10.1515/ling.2007.009. https://doi.org/10.1515/ling.2007.009
  2. Beths, Frank. 1999. The history of dare and the status of unidirectionality. Linguistics 37(6). DOI: 10.1515/ling.37.6.1069. https://doi.org/10.1515/ling.37.6.1069
  3. Lass, Roger. 1990. How to do things with junk: exaptation in language evolution. Journal of Linguistics 26(1): 79–102. DOI: 10.1017/S0022226700014432. https://doi.org/10.1017/S0022226700014432
  4. Haspelmath, Martin. 1999. Why is grammaticalization irreversible? Linguistics 37(6). DOI: 10.1515/ling.37.6.1043. https://doi.org/10.1515/ling.37.6.1043
  5. Campbell, Lyle. 2000. What's wrong with grammaticalization? Language Sciences 23(2–3): 113–161. DOI: 10.1016/S0388-0001(00)00019-X. https://doi.org/10.1016/S0388-0001(00)00019-X
  6. Newmeyer, Frederick J. 2000. Deconstructing grammaticalization. Language Sciences 23(2–3): 187–229. DOI: 10.1016/S0388-0001(00)00021-8. https://doi.org/10.1016/S0388-0001(00)00021-8
  7. Janda, Richard D. 2000. Beyond 'pathways' and 'unidirectionality': on the discontinuity of language transmission and the counterability of grammaticalization. Language Sciences 23(2–3): 265–340. DOI: 10.1016/S0388-0001(00)00023-1. https://doi.org/10.1016/S0388-0001(00)00023-1
  8. Heine, Bernd, and Tania Kuteva. 2003. On contact-induced grammaticalization. Studies in Language 27(3): 529–572. DOI: 10.1075/sl.27.3.04hei. https://doi.org/10.1075/sl.27.3.04hei
  9. Norde, Muriel. 2009. Degrammaticalization. Oxford: Oxford University Press. DOI: 10.1093/acprof:oso/9780199207923.001.0001. https://doi.org/10.1093/acprof:oso/9780199207923.001.0001
  10. Willis, David. 2010. Degrammaticalization and obsolescent morphology: Evidence from Slavonic. Studies in Language Companion Series 119: 151–178. DOI: 10.1075/slcs.119.09wil. https://doi.org/10.1075/slcs.119.09wil
  11. Trousdale, Graeme, and Muriel Norde. 2013. Degrammaticalization and constructionalization: two case studies. Language Sciences 36: 32–46. DOI: 10.1016/j.langsci.2012.03.018. https://doi.org/10.1016/j.langsci.2012.03.018
  12. Norde, Muriel, and Karin Beijering. 2014. Facing interfaces: A clustering approach to grammaticalization and related changes. Folia Linguistica 48(2): 385–424. DOI: 10.1515/flin.2014.014. https://doi.org/10.1515/flin.2014.014 (open-access copy: https://edoc.hu-berlin.de/bitstreams/7d6e6118-3adf-4938-af05-219148def65e/download)
  13. Viti, Carlotta. 2015. On degrammaticalization: Controversial points and possible explanations. Folia Linguistica 49(2). DOI: 10.1515/flin-2015-0014. https://doi.org/10.1515/flin-2015-0014
  14. Franco, Ludovico. 2015. Axial Parts, Phi-Features and Degrammaticalization: the Case of Italian Presso/Pressi in Diachrony. Transactions of the Philological Society 114(2): 149–170. DOI: 10.1111/1467-968X.12067. https://doi.org/10.1111/1467-968X.12067
  15. D'Alessandro, Roberta. 2014. Death and contact-induced rebirth of impersonal pronouns. A case study. Probus 26(2): 249–274. DOI: 10.1515/probus-2014-0009. https://doi.org/10.1515/probus-2014-0009
  16. Willis, David. 2016. Exaptation and degrammaticalization within an acquisition-based model of abductive reanalysis. In Exaptation and Language Change (Current Issues in Linguistic Theory 336): 197–225. DOI: 10.1075/cilt.336.07wil. https://doi.org/10.1075/cilt.336.07wil
  17. Van Goethem, Kristel, and Muriel Norde. 2020. Extravagant 'fake' morphemes in Dutch. Morphological productivity, semantic profiles and categorical flexibility. Corpus Linguistics and Linguistic Theory 16(3): 425–458. DOI: 10.1515/cllt-2020-0024. https://doi.org/10.1515/cllt-2020-0024
  18. Herda, Damian. 2023. From Bound to Free: On the Degrammaticalization of the Polish Intensifying Prefix prze-. Scando-Slavica 69(2): 161–178. DOI: 10.1080/00806765.2023.2273332. https://doi.org/10.1080/00806765.2023.2273332
  19. Company Company, Concepción. 2024. Grammaticalization and Diachronic Structural Pattern. Catalan Journal of Linguistics 23: 13–39. DOI: 10.5565/rev/catjl.460. https://doi.org/10.5565/rev/catjl.460
  20. Echeverría, Carlos I. 2026. Again on the Existence of Causative Periphrases in Spanish: The Case of 'ENVIAR/MANDAR a + Infinitive'. Languages 11(5): 90. DOI: 10.3390/languages11050090. https://doi.org/10.3390/languages11050090