Literal and metaphorical meaning: In search of a lost distinction

Nicholas Allott

Intuitively, there is a distinction between (1a–c) and 1d: the former are literal uses; the latter is metaphorical:

(1a) cut bread
(1b) cut hair
(1c) cut the grass
(1d) cut taxes

On one traditional view, metaphorical uses are ones that express a different sense from the literal i.e. linguistic i.e. encoded sense.

But it seems to be false that each word has one sense which is expressed in every literal use: the senses of cut in each of (1a–c) seem to be different. Similarly (2a–b) show two literal senses of book:

(2a) John read the book on his computer [book content]
(2b) John burned the book [book object]

So the traditional criterion won’t distinguish literal and metaphorical uses.

Could book, cut etc. be homonyms? No: there is recent experimental evidence that the distinction between homonyms (e.g. bat [rodent] and bat [sports]) and polysemes (e.g. different senses of the noun book) is cognitively real: they are processed differently. In particular, in processing polysemes, there’s no bias towards the most frequent or otherwise dominant sense (Frisson, 2015), while it has long been known that there is such a bias for homonyms.

It seems, therefore, that the linguistic meaning of a polysemous word is something that is neutral between its senses and not equal to any of them. So the criterion above for figurative meaning is a non-starter.

Another possible criterion is that metaphorical uses are those which drop a central feature of the literal sense (Allott & Textor, 2017). E.g. <affects material object> in (1d) and <fish> in (3):

(3) My lawyer is a shark.

But cross-linguistic work (Glanzberg, 2008; Spalek, 2015) shows that at least some syntacto-semantic features of words persist in metaphorical uses. Compare English cut with Spanish cortar. cortar has some kind of aspectual feature (rather like English destroy), which makes it inapplicable to cases where merely some part is removed (as in English cut away). Crucially this bars both literal uses which are not cases of (roughly) irreversible and total severance and metaphorical extensions like (the Spanish version of) cut taxes by 20%. (Cf. English *destroy taxes by 20% vs. diminish taxes by 20%.)

Where does this leave the search for a criterion for metaphorical uses? I propose that i) we should distinguish between linguistic features and conceptual features, ii) in metaphorical use at least one conceptual feature is dropped; iii) linguistic features are never dropped.

I attempt to support (or refute) these claims by a preliminary survey of relavant examples (e.g. of uses of cut) that can be found in corpora.



Allott, N. & Textor, M. (2017). Lexical modulation without concepts: Introducing the derivation proposal. Dialectica, 71(3), 399–424.

Frisson, S. (2015). About bound and scary books: The processing of book polysemies. Lingua, 157, 17–35.

Glanzberg, M. (2008). Metaphor and lexical semantics. The Baltic International Yearbook of Cognition, Logic and Communication, 3, 1–47.

Spalek, A. A. (2015). Spanish change of state verbs in composition with atypical theme arguments: Clarifying the meaning shifts. Lingua, 157, 36–53.

Publisert 16. nov. 2019 16:29 - Sist endret 16. nov. 2019 16:29