Show Summary Details

Page of

Printed from Oxford Classical Dictionary. Under the terms of the licence agreement, an individual user may print out a single article for personal use (for details see Privacy Policy and Legal Notice).

date: 01 March 2021


  • Sebastian Matzner


The term metonymy denotes a literary trope, that is, a specific form of defamiliarized expression, which indirectly refers to what is at issue. Metonymy achieves this by way of exploiting an already existing association between the term (or terms) used metonymically—the metonym—and the term (or terms) implicitly at issue. Metonymy thus differs from metaphor, among other things, in that it does not invoke an underlying analogy or similarity between what is said and what is at issue. In both ancient and modern criticism, metaphor received significantly more attention than metonymy (partly owed to the fact that the poetic effects of metaphor tend to eclipse those of metonymy, partly because of the stronger appeal of the logical dimension at the heart of metaphor). As a result, metonymy—though widely used—is often ill-defined as a critical concept. Today, it features in literary-aesthetic, diachronic-etymological, (post-)structuralist, and cognitive criticism. Ancient literature, both Greek and Latin, is rich in metonymic usages, albeit with varying degrees of poetic intensity; the pattern is one of relatively few intense outcomes, and relatively many less intense ones. Prominent among the general literary-aesthetic effects of metonymy’s semantic shifts are the creation of a poignantly condensed impression of what is at issue; a change in focalization by zooming out onto a higher plane or zooming in on newly foregrounded micro-level aspects; and movement between the concrete-material and the abstract-conceptual dimensions of what is at issue.

The term metonymy denotes a form of defamiliarized expression, which indirectly refers to what is at issue by way of exploiting an already existing association between the term (or terms) used—the metonym—and the term (or terms) implicitly at issue. Metonymy differs from metaphor, among other things, in that it does not invoke an underlying analogy or similarity between what is said and what is at issue. Although now customarily seen as the two most central tropes, metaphor and metonymy emerge onto the scene of rhetoric and poetics on a markedly unequal footing—and despite metonymy’s gradual rise to greater prominence in literary criticism (and beyond), especially during the 20th century, critical discussion and aesthetic appreciation of metaphor tend to eclipse those of metonymy (see rhetoric, Greek and rhetoric, Latin).

The roots of this imbalance can be traced back to Aristotle. The philosopher not only privileges metaphor by singling it out repeatedly for special praise on account of its logical base in the principle of analogy and similarity (Poet. 1459a5–8; Rh. 1405a8–13, 1410b17–20); he also uses the same technical term, μεταφορά‎, in the narrow sense, to refer to metaphor as a particular trope, characterized by the analogy principle, and as an umbrella term for tropes in general, encompassing every “transference of a term from one thing to another” (Poet. 1457b7–32), thereby subsuming, as rightly noted by Cicero (Orat. 94), instances that include dead metaphor (νηũς δέ μοι ἥ δ‎’ ἕστηκεν, “here stands my ship”), metonymy (τεμὼν ταναήκεϊ χαλκῷ‎, “cutting off with the slender-edged bronze”), and metaphor proper (δυσμὰςβίου‎, “life’s sunset”). With metonymy thus left undifferentiated, undiscussed, and undefined, the subsequent rhetorical tradition has witnessed centuries of uncertainty and debate concerning metonymy’s core principle and its relationship to metaphor (as either a variety of it or as different in kind) and to other tropes (especially synecdoche).

Ancient criticism uses a variety of terms to talk about metonymy: alongside μετωνυμία‎—first attested in Tryphon (2) (Trop. 739, 20–1) but used there as an already established term—and the latinized metonymia, one also finds the calque denominatio (Rhet. Her. 4.43) and immutatio/παλλαγή‎ (Cic. Orat. 93). Whatever the term used, the attempted explanations are typically vague and largely centre on the notion of a substitution of terms, without clarifying what principle enables said substitution. Quintilian’s definition of metonymy—“which is the substitution of one name for another” (metonymia, quae est nominis pro nomine positio; Quint. Inst. 8.6.23)—is representative and, much like Aristotle’s problematically broad notion of μεταφορά‎, it covers, in fact, all tropes in general rather than describing metonymy as one particular case. Here, as in the Greek grammarians and rhetoricians, and throughout the late antique, medieval, and early modern rhetorical tradition, it is only the list of examples given after such a “definition” that conveys some sense of the specific phenomenon under discussion. Such lists typically contain a series of clichéd stock metonyms, which are grouped in broad conceptual categories (such as “effect for cause,” “container for contained,” “producer for product,” “deity for remit,” “material for object,” “place for inhabitant,” and so on); yet the arbitrariness of the selected categories and their potential limitlessness only bespeak the difficulty of formulating one structural principle behind metonymy’s (in conceptual terms) multifarious manifestations. Cicero’s discussion of metonymy—in marked contrast to his treatment of metaphor—illustrates the point: having defined metaphorical expressions (tralata uerba) as operative on the basis of a perceived or construed similarity (per similitudinem), he characterizes metonymic expressions (immutata) much more loosely as those “in which a proper word is replaced by another proper word, which means the same but is taken from some other matter that follows on [from what is at issue]” (in quibus pro uerbo proprio subicitur aliud quod idem significet sumptum ex re aliqua consequenti; Cic. Orat. 92). The notion of replacing a proper word with another proper word (cf. also De or. 3.167) is strikingly paradoxical: after all, any trope is an instance of defamiliarized language that noticeably deviates from ordinary usage and therefore involves, by definition, the im-proper use of words (which makes tropes perceptible and poetically effective in the first place). Cicero’s formulation and impression are, however, indicative of metonymy’s frequent association with prose, ordinary language, and a (relatively) low degree of poeticity. At the same time, Cicero’s description of the relationship between the “substituting” and the “substituted” element in metonymy as something “following on from” or “accompanying” what is at issue (consequens)—much like the Auctor ad Herennium’s similar characterization of metonymy as an expression that draws from a “near”/“neighbouring” (propinquus) or “bordering”/“adjoining” (finitimus) matter (Rhet. Her. 4.43)points to the modern term widely used wherever definitions of metonymy are attempted: contiguity.

In the 20th century, metonymy was popularized by Roman Jakobson and began to be seen as the polar opposite of metaphor: in the context of his influential widening of metaphor and metonymy—from specific modes of language usage to universal structural-semiotic patterns that can be found in any given sign system (both verbal and non-verbal)—the term contiguity promises to furnish metonymy with a neat equivalent to metaphor’s characteristic principle of analogy. In practice, however, what is concretely meant by contiguity often remains elusive and underdetermined: covering all sorts of relations (except those of analogy and similarity), metonymic and contiguity become highly malleable placeholders, chiefly determined negatively in their specific context, namely through counter-distinction from their respective “metaphorical” opposites, but with little stable positive meaning of their own. Jakobson’s suggestive claim that metaphor characterized lyric poetry, Romanticism, Chaplin’s films, and Freudian dream symbols, while metonymy was embodied in epic poetry, realist novels, Griffith’s films, and Freudian dream projections, and triggered structuralist explorations of metonymic phenomena in a wide range of fields; yet, irrespective of the individual insights yielded by such studies, collectively they prove barely relatable to a shared notion of metonymy as a common point of reference. The blurred understanding of metonymy that results is compounded further by the recurrent failure to distinguish clearly between different areas of study which invoke metonymy in markedly different ways and with differently geared interests. Four broad perspectives can be distinguished.


Literary-aesthetic: concerned with stylized, poetically effective language usage that deviates from ordinary language usage; analysis typically concentrates on expressions and sentences (and, crucially, their aesthetic and semantic impact) by invoking metonymy as a particular type of deviation, sc. a literary trope.


Diachronic-etymological: focused on the historical development of ordinary language; analysis typically concentrates on the semantic scope of individual words, with metonymy invoked as one pattern (among others) to describe the gradual historical change or extension of a word’s meanings in ordinary usage, often to explain the origin of word’s polysemy in ordinary usage.


Structuralist: interested in broader underlying frameworks and arrangements that facilitate and shape the generation of meaning in cultural products; analysis here centres on larger scale patterns and implicit organising principles, with appeals to metonymy referencing a manner in which individual elements of a sign system relate to others (for instance: metonymic cuts that relate one scene in a film to another, or metonymic plays or narratives whose dramaturgy or plot progression follows a pattern associated with contiguity as the principle of metonymy).


Cognitive: chiefly directed at the conceptual work of expressions that operate along the lines of metonymy, regardless of whether they occur in conventional, ordinary language or as poetic usage in marked deviation from it; analysis typically concentrates on mapping the domains involved and exploring the implications for the conceptualization and processing of content.

While all four of these can and do intersect, problems occur when they are conflated. From a diachronic-etymological perspective, for instance, the semantic scope of robur (“oak-tree,” “strength”) can be explained as resulting historically from metonymic shifts originating in the root *rob- (“red”) that characterize the sturdiest part of this tree’s trunk; or πορφύρα‎, which covers the “purple-fish,” “purple dye” (obtained from it), and a “purple garment” (dyed with it) as emerging by way of gradual metonymic extension. Yet it would be misleading to consider the occurrence and use of either term in any of these meanings to be metonymic from a literary-aesthetic perspective, since these extended meanings, once incorporated into the terms’ denotative meaning, cease to be poetically effective as non-ordinary, deviant usage (even if, in some cases, they may still be recognizable as secondary rather than primary meanings). Similarly, clichéd metonymic expressions that have entered common parlance are often studied with respect to their conceptual domains from a cognitive perspective, but hold limited literary-aesthetic significance, just as a play whose dramaturgy could be described in structuralist terms as unfolding in a metonymic manner may contain few or no metonyms but abound in metaphors as far as its poetic language is concerned.

If these problems trouble discussions of metonymy in modern criticism, literary criticism in antiquity is marked by a different set of issues whose impact—mediated through the rhetorical handbook tradition—can be felt to this day. Given their strong pragmatic orientation towards guidance for best practice (either alongside or in lieu of abstract theorization), the ancient critics are mostly concerned with the appropriate and effective use of metonymic expressions, usually treated among other means of stylization and considered relevant to both poetics and rhetoric—verse and prose (the recurrent advice on restrained use of any kind of intensely defamiliarized language in classical rhetoric notwithstanding, cf. e.g., Arist. Rh. 1404a20-1404b25, 1405a2-8; Cic. Orat. 201-2; Quint. Inst. 8.6.17-20, 24-25). Explicit discussion of the aesthetic effect of metonymy, however, is limited and hampered further by the pervasive use of clichéd stock examples. Quintilian (Inst. 8.6.24) admits as much when noting that referring to fire through its presiding deity “Vulcan” is a matter of common speech (“Vulcanum” pro igne vulgo audimus), and his assessment of structurally equivalent expressions—like “having fought with varying Mars [sc. success in battle]” (“vario Marte pugnatum”) as more learned (eruditus est sermo) and “Venus” for “sexual intercourse” as more socially acceptable (“Venerem” quam coitum magis decet)—says little about the particulars of metonymy’s characteristic effects. Cicero’s (de Orat. 3. 167-8) discussion is representative in this respect: having likewise listed only clichéd expressions for illustrative purposes (e.g., “curiam” pro senatu and “tela” pro bello, “‘the curia’ for the Senate” and “‘spears’ for war”), he endorsed frequent use of this device (saepe sumendus) but failed to say anything specific about the nature of its praised embellishing effect (grauis est modus in ornatu orationis; commutato uerbo res eadem enuntiatur ornantius).

To fully appreciate the rich and varied contributions that metonymic expressions make in literary-aesthetic terms to the texture of Greek and Latin texts, metonymy’s core principle of contiguity is best approached as lexical contiguity: a metonymically used term (the metonymic vehicle) is recognized as such because it cannot be read literally in its given context but gestures beyond itself to what is at issue (the metonymic tenor). What is at issue is ascertained by way of inferring a suitable term or terms—not along the lines of some implied analogy or similarity (as in metaphor), but from out of the semantic field referenced by the present metonymic vehicle. Such a semantic field can be defined pragmatically as the sum total of all other words any given word can be—and is—collocated within ordinary language usage. Thus, in the description of Ismarus in Vergil’s Aeneid (10.140) as “aiming wounds and arming shafts with venom” (vulnera derigere et calamos armare veneno), vulnera cannot be taken in its ordinary meaning of “wounds” (given the incompatibility of this meaning with its status as object of derigere, “to aim”); instead, it references a semantic field, within which—guided by the demands of the present context—terms like weapon, missile or arrow come into view as viable metonymic tenor(s). The operative basis for metonymic inference is thus the collocatability of metonymic vehicle and tenor, that is, their ability to occur together in a coherent ordinary language sentence (here, for example, derigere sagittas ad infligendum vulnera, “aiming arrows so as to inflict wounds”). Irrespective of any underpinning logic, metonymy’s signature trait thus consists in a noticeable, aesthetically effective lateral shift within the terminology of one semantic field.

While it may be tempting to consider the above instance of metonymy—in the spirit of the rhetorical handbook tradition—a simple substitution of “cause for effect,” the assumption that metonyms are underwritten by a straightforward logical relationship does not bear scrutiny. The limits of this approach can be seen in this verse from Aeschylus’ Agamemnon, where the Greek victory over Troy is prophesied with the words “In time this path/journey will capture the city of Priam,” χρόνῳ μὲν ἀγρεῖ Πριάμου πόλιν ἅδεκέλευθος‎ (126). Again, the metonymic vehicle, “path/journey” (κέλευθος‎), is marked as a trope by its incompatibility with the main verb, “will capture” (ἀγρεῖ‎), which calls for human agents, so that from out of the semantic field referenced (by the metonymic vehicle), terms like military expedition, army, or soldiers come into view as metonymic tenor(s)—yet one would struggle to formulate any abstract logical principle to describe the relation between vehicle and tenor(s) beyond offering an expanded, explanatory sentence like “The army’s journey to Troy will lead to the city’s capture,” which exemplifies, precisely, their collocatability in ordinary language, which establishes the lexical contiguity that enables metonymic usage.

The above definition, focused on terms and terminology rather conceptual domains and logical relations, not only has the advantage of capturing all manifestations of metonymy (different though they may be in logical-conceptual terms) in one principle. It moreover explains:


Why metonymy is often associated with prose, ordinary language, and (relatively) weak poetic effects—because it draws on pre-existing relations between terms, created in and through ordinary language usage, which limits its potential for powerful deviation from it;


How metonymy differs from metaphor—in that it is marked by a shift within one semantic field, rather than staging within itself a clash of two different semantic fields as metaphor’s analogy principle does (compare and contrast from Aristotle’s examples the metonym τεμὼν ταναήκεϊχαλκῷ‎, “cutting off with the slender-edged bronze,” with the metaphor δυσμὰςβίου‎, “life’s sunset”);


How it relates to other tropes—notably to synecdoche, which, in keeping with Quintilian’s accurate observation that there is no great gap between them (nec procul ab hoc genere [sc. synecdoche] procedit μετωνυμία‎; Inst. 8.6.23), emerges as a subcategory of metonymy, since it is equally based on lexical contiguity through collocation in ordinary usage (here established specifically through utterances concerning the composition of wholes by parts), just as personification is a subcategory of metaphor in which the general analogy-principle is realized specifically by suggesting an analogy to human agency.

From the fundamental difference in their governing core principles follows a difference in the degree of poetic intensity characteristically generated by metaphors and metonyms, with analogy facilitating more remarkable terminological clashes in metaphor, and lexical contiguity facilitating less remarkable intra-terminological shifts in metonymy. The preponderance of clichéd expressions, firmly established early on in both Greek and Latin literary idiom (e.g., κάρα‎/κεφαλή‎/caput, “head,” referring to a person; στέγη‎/tectum, “roof,” for “house”; σκάφος‎/carina, “keel,” for “ship”; and many more), adds further to the perception of metonymy as primarily a matter of the elevated register of poetic diction rather than of strikingly heightened poetic speech. While this underscores metonymy’s important role in formal stylization (and classicizing style), it is apparent that metonymy, too, can produce striking effects, even if its overwhelming usage is indeed of a refined or restricted kind: the general pattern is one of relatively few intense outcomes and relatively many less intense ones. Where intense effects occur, they are typically located not within the metonym itself (owing to the relative terminological coherence and pre-established connection of the metonymic shift itself), but at the interface of the metonym and its context. This includes in particular scenarios where the metonymic shift creates arresting juxtapositions, pronounced abrasiveness, or other poignant interactions between the metonymically used term and terms of the immediate context (e.g., Ov. Am. 1.15.15nulla Sophocleo veniet iactura cothurno, “no loss of stature will come to the Sophoclean buskin,” invoking tragedy via the thick-soled boots worn by tragic actors, and contrasting fears of diminishment with the height-adding footwear); or where the surrounding context leaves the metonym markedly underdetermined so that a particularly wide range of possible metonymic tenors is kept simultaneously at play (e.g., Pind. Pyth. 10.36 γελᾷθ᾽ὁρῶνὕβρινὀρθίανκνωδάλων‎, “he laughs to see the beasts’ upright/high-pitched transgression,” pointing to the donkeys’ excessive jumping, braying, and/or erections); or where the metonymically used term resonates disproportionately strongly and significantly with a text’s wider networks of imagery (i.e., the metonym facilitates important imagistic connections, as between Aesch. Ag. 239κρόκου βαφὰςδ᾽ἐς πέδον χέουσα‎, “as she poured saffron dye toward the ground,” pointing to Iphigenia’s saffron-coloured garment right before her violent death, and 958-60 στιν θἀλασσαν . . . τρέφουσα πολλῆς πορφύρας . . . κηκῖδα‎ . . . εἱμάτωνβαφάς, “there is the sea . . . nourishing much purple . . . that gushes forth, a dye for clothing,” pointing to the carpet that leads Agamemnon to his own) or even concentratedly captures a text’s key concerns (i.e., the metonym reads as a mise en abyme, as in Verg. Aen. 1.1: arma uirumque cano, “I sing of arms and the man”).

Prominent among the more general literary-aesthetic effects of metonymy’s lateral shifts are the creation of a poignantly condensed impression of what is at issue (as in the Vergilian example above, where the effect is one of bringing into focus the virtually instant impact of Ismarus’ battlefield skills, portraying him as directly distributing wounds, thereby eliding the intervening distance of both the time lapsed and the weapon used as well as eliminating any scope for missed targets); a change in focalization by zooming out onto a higher plane or zooming in on newly foregrounded micro-level aspects (e.g., Pind. Pyth. 9.80-1 Εὐρυσθῆος ἐπεὶ κεφαλὰν | ἔπραθε φασγάνου ἀκμᾷ‎, “after he cut off Eurystheus’ head with the edge of his sword”); movement between the concrete-material and the abstract-conceptual dimensions of what is at issue (e.g., Pind. Pyth. 5.34 κατέκλασε γὰρ ἐντέων σθένος οὐδέν‎, “For in no way did he break the strength of the equipment”; Cic. Off. 1.77cedant arma togae, concedat laurea laudi (“weapons [sc. war] shall give way to the toga [sc. civic life in peace-time] and laurels [sc. military triumph] to praises [sc. civic eulogies]”).

Whether metonymy rather than metaphor and simile is relatively more prominent in Latin literature—on its own or in comparison to Greek literature—is a question for which only a large-scale evaluation will be able to offer an answer. Notable instances of metonymy in canonical Latin works combined with the views relayed by Latin writers themselves regarding the supposed “poverty” of their language and the explicit acknowledgment of a concomitant need for its creative expansion may foster this impression (cf. e.g., Lucr. 1.136-9, 830-2, 3.258-60; Cic. Fin. 1.10, 3.5; Sen. Ep. 58.1, 7; Plin. Ep. 4.18.1; but note here, too, the issue of poetic/live metonymy versus clichéd/dead metonymy, i.e., the difference between literary-aesthetic and diachronic-etymological perspectives); however, an empirical evaluation of (at least a representative selection of) the extent texts would be required to corroborate any such impression.


  • Allan, Kathryn. Metaphor and Metonymy: A Diachronic Approach. Chichester, UK: Wiley-Blackwell, 2008.
  • Bredin, Hugh. “Metonymy.” Poetics 5, no. 1 (1984): 45–58.
  • Coleman, Robert. “Poetic Diction, Poetic Discourse, and the Poetic Register.” In Aspects of the Language of Latin Poetry. Edited by James N. Adams and Roland G. Mayer, 21–96. Oxford: Oxford University Press, 1999.
  • Ferri, Robert. “The Language of Latin Epic and Lyric Poetry.” In A Companion to the Latin Language. Edited by James Clackson, 344–366. Chichester, UK: Wiley-Blackwell, 2011.
  • Jakobson, Roman. “Two Aspects of Language and Two Types of Aphasic Disturbances.” In Roman Jakobson: Selected Writings: Word and Language, Vol. 2. Edited by Stephen Ruby, 239–259. The Hague: Mouton, 1971.
  • Landfester, Manfred. Einführung in die Stilistik der griechischen und lateinischen Literatursprachen. Darmstadt, Netherlands: Wissenschaftliche Buchgesellschaft, 1997.
  • Lausberg, Heinrich. Handbuch der literarischen Rhetorik. Eine Grundlegung der Literaturwissenschaft (3rd ed.). Stuttgart: Steiner, 1990. = Handbook of Literary Rhetoric: A Foundation for Literary Studies. Edited by R. D. Anderson and D. E. Orten, translated by M. T. Bliss, A. Jansen, and D. E. Orten: §§565–571. Leiden, The Netherlands: Brill, 1998.
  • Lodge, David. Modes of Modern Writing: Metaphor, Metonymy, and the Typology of Modern Literature. London: Edward Arnold, 1977.
  • Matzner, Sebastian. Rethinking Metonymy: Literary Theory and Poetic Practice from Pindar to Jakobson. Oxford: Oxford University Press, 2016.
  • Nagy, Gregory. Masterpieces of Metonymy: From Ancient Greek Times to Now. Hellenic Studies Series 72. Washington, DC: Center for Hellenic Studies, 2015.
  • Panther, Klaus-Uwe, and Günter Radden, eds. Metonymy in Language and Thought. Amsterdam: Benjamins, 1999.
  • Silk, Michael S. “Metaphor and Metonymy. Aristotle, Jakobson, Ricoeur, and Other.” In Metaphor, Allegory, and the Classical Tradition: Ancient Thought and Modern Revisions. Edited by George R. Boys-Stones, 15–47. Oxford: Oxford University Press, 2003.
  • Silk, Michael S. “The Language of Greek Lyric Poetry.” In A Companion to the Ancient Greek Language. Edited by Egbert J. Bakker, 424–440. Oxford: Wiley-Blackwell, 2010.
  • Steen, Gerard. “Metonymy Goes Cognitive-Linguistic.” Style 39, no. 1 (2005): 1–11.