Linguistic Metonymy: Implicitness and Co-Activation of Mental Content

1Introduction

1.1 The pre-cognitive linguistic tradition

For centuries, metonymy and metaphor have been studied as tropes and as figures of speech, and as such they used to be studied primarily within the context of rhetoric. Their position, function, and importance within the system of tropes have been in constant transition. Aristotle, in his Poetics, uses the term ‘metaphor’ as an umbrella term for everything that is not named by its own name (1457b1–1458a17). In his Rhetoric, he distinguishes four types of metaphor. Among these it is analogy-based metaphor that comes closest to current definitions of metaphor (1410b–1412b). Traditional rhetoric has set up numerous subclasses of Aristotle’s notion of metaphor. As Benczik (2005) points out, Cornificius distinguishes ten, and Quintilian fourteen, tropes. Based on classic works of rhetoric, Lausberg (1990) lists nine tropes. The aim of traditional rhetoric in creating these refined distinctions was practical: it served the purpose of educating rhetors.

The plurality of sub-types of tropes was reduced by Burke (1945), based on the relationship between them, to four major types: metaphor, metonymy, synecdoche, and irony. Synecdoche is often considered to be a subtype of metonymy, whereas according to Lausberg, irony can be traced back to metaphor:¹ “Die Ironie ist der contrarium-Grad der Metapher” (Lausberg 1990: 303 cited in Benczik 2005). Consequently, the two remaining major types of tropes were metaphor and metonymy.²

Structural linguistics was the first school of linguistics which focused on the linguistically oriented study of metaphor and metonymy. In Jakobson’s view (2002), the two phenomena are completely distinct from each other. Metaphor is based on similarity, i.e. on intrinsic properties, whereas metonymy on contiguity, i.e. on extrinsic properties. According to Jakobson, metaphor is a phenomenon of the paradigmatic pole of language which involves the operation of selection, ←11 | 12→unlike metonymy, which is to be located on the syntagmatic pole of language involving the operation of combination.

Unlike structural approaches to metaphor and metonymy, which draw sharp boundaries between the two phenomena, later theories emphasize the relation between them, primarily the metonymic motivation and basis of metaphor (e.g. Eco 1971, Kemény 2002, and Benczik 2005). However, the linguistic study of metonymy and metaphor, and the relation between them, was revolutionized by the appearance of Lakoff and Johnson’s seminal work Metaphors We Live By in 1980, marking the beginnings of the holistic cognitive linguistic study of metonymy and metaphor. Influenced by the state of the art results of cognitive psychology, the most important claim of their theory is that metonymy and metaphor are not merely linguistic ornaments, or figures of speech, but are fundamental operations in human cognition, i.e. they are figures of thought.³

1.2From the poor sister to ubiquity

In the holistic cognitive linguistic literature on metonymy, it is a very common observation that the interest devoted to metonymy shrinks into insignificance beside the attention directed towards metaphor. For a long time, metonymy was looked upon as “metaphor’s poor sister”⁴ and metaphor was thought of as “metonymy’s rich relative”⁵. A closer inspection of holistic cognitive linguistic works on conceptual metaphor and conceptual metonymy does indeed show that the contemporary research on metonymy has been developing in the shadow of the research on metaphor. As a result of this asymmetric interest, the holistic cognitive linguistic research on metonymy lags behind the intensive investigations into metaphor, while the cognitive investigations into metonymy generally run in the same direction.

The relatively late and initially modest interest in metonymy is also indicated by the fact that Lakoff and Johnson (1980) devote merely a single chapter to metonymy (Chapter 9), whereas the first collection of papers primarily concerned with metonymy only appeared almost twenty years later in Günter Radden and Klaus-Uwe Panther’s edition (1999). ←12 | 13→

During these initial two decades of holistic cognitive linguistics, metonymy had suffered from a relative lack of interest and had fallen behind in comparison with the results of metaphor research. For instance, in the case of metonymy it is rather difficult to outline a relatively unified standard theory and its later improved, developed, or more elaborate versions, as can be done with metaphor (Lakoff and Johnson 1980, Lakoff 1987, 1993, Lakoff and Johnson 1999; the theory of primary metaphors by Grady 1997a and 1997b; or the neural theory of metaphor as proposed by Lakoff 2008 or Feldman 2006). Furthermore, elaborate rival theories of metonymy outside the framework of holistic cognitive linguistics hardly exist, while in metaphor research they are clearly present, as for instance Glucksberg’s and his colleagues’ property attribution theory of metaphor (e.g. Glucksberg and Keysar 1990, 1993, Glucksberg, McGlone and Manfredi 1997, Glucksberg 2001, 2003), Gentner and her colleagues’ structure mapping theory of metaphor (e.g. Gentner 1983, Gentner and Toupin 1986, Bowdle and Gentner 2005, Gentner and Bowdle 2001, 2008, Gentner et al. 2001 etc.), or relevance theoretic approaches to metaphor (e.g. Wilson and Carston 2006, Carston 2010, Sperber and Wilson 2008, Wilson and Sperber 2012).⁶

Moreover, cognitive research on metonymy also shows a deficit of empirical foundations when compared to the massive amount of empirical data accumulated in the cognitive research on metaphor. Generally accepted and practiced corpus linguistic methods of metonymy research are lacking; though broad cross-linguistic investigations into metonymy seem to be a promising and fruitful venture, they are still in their infancy; and finally, the experimental examination of metonymy is almost completely absent. Whereas metaphor is nowadays often examined with the help of cutting-edge methods such as eye-trackers, functional neuroimaging procedures (fMRI), and event-related brain potentials (ERPs) (e.g. Forgács et al. 2012, Forgács et al. 2014, Forgács 2014 and the literature reviewed there, or Forgács et al. 2015), metonymy is predominantly investigated by intuitive-introspective and manual, small-scale corpus methods.

Another facet of cognitive metonymy research as it developed in the shadow of cognitive metaphor research is that initially, the primary concern of metonymy researchers was the problems of metaphor-metonymy distinction and the cognitive and linguistic interaction between the two phenomena (e.g. Barcelona 2000a, Dirven and Pörings 2002), while since then, the main objective of metonymy research has gradually become to point out and to support the ←13 | 14→ubiquity of conceptual metonymy and its primacy relative to conceptual metaphor in human thinking and reasoning and in natural language (e.g. Panther and Radden 1999a, Panther and Thornburg 2003a).

As a result of these endeavors, in holistic cognitive linguistics metonymy is now generally considered as a general cognitive mechanism that plays a central and even more fundamental role in every field of conceptual and linguistic organization than metaphor. Accordingly, the study of metonymy as a linguistically manifest phenomenon has been pushed into the background, something which is clearly indicated by the lack of a generally accepted and functional distinction between conceptual metonymy and metonymically motivated expressions, and by the fact that the relationships between these are rather vaguely formulated. Metonymic expressions are most often considered to be simply the manifestations of conceptual metonymies on the level of linguistic units. However, if we accept the ubiquity view of conceptual metonymy in its broadest form, this implies that language is essentially metonymic, which in turn leads to the conclusion that each and every linguistic expression is metonymic. This conclusion may well result in the unfortunate situation that the notion of ‘metonymy’ and especially that of ‘metonymic expression’ might become entirely limitless unless they are defined with the help of linguistically manifest properties in addition to their general conceptual metonymic motivation.

1.3Problems under scrutiny

As a result of metonymy’s evolution from metaphor’s poor sister to a ubiquitous conceptual phenomenon, the notion of ‘metonymy’ as it is generally accepted within holistic cognitive linguistics runs the risk of becoming indefinite and unlimited. This risk can be best grasped in the form of two interrelated problems, whose solution may contribute to its elimination. First, the range of metonymic phenomena is hard to distinguish from, and to delimit against, other related and similar phenomena, such as metaphor and so-called active zone phenomena. Here we face a definitional and distinction problem, which can be formulated as follows:

The problem of delimitation (P_DEL): On the basis of what criteria can metonymy be delimited against related phenomena?

Second, the notion of metonymy embraces such a broad range of heterogeneous phenomena that the possibility of formulating generalizations which are valid for the whole set of metonymic phenomena becomes very limited. In other words, the second problem is of a classificatory nature: ←14 | 15→

1.4Metonymy in holistic cognitive linguistics

Before outlining my approach to (P_DEL) and (P_CLASS) it is worth taking a brief look at what is understood generally by metonymy in holistic cognitive linguistics. The most reasonable starting point to illustrate the holistic cognitive linguistic notion of metonymy is to quote one of the broadest and most generally accepted definitions, that proposed by Günter Radden and Zoltán Kövecses, who laid down the foundations of a theory of conceptual metonymy (Kövecses and Radden 1998, Radden and Kövecses 1999: 21): “Metonymy is a cognitive process in which one conceptual entity, the vehicle [source], provides mental access to another conceptual entity, the target, within the same idealized cognitive model.”

To gain a better understanding of the definition, we need to consider what it means to provide mental access and what idealized cognitive models are. In Langacker’s approach (1993, 1999) metonymic expressions function as reference or access points to mental content. They open up a chunk of structured knowledge, within which the reference point (or source) makes the target mentally available for meaning construction purposes: “The entity that is normally designated by a metonymic expression serves as a reference point affording mental access to the desired target (that is, the entity actually being referred to)” (Langacker 1993: 30).

The knowledge structure to which both the explicitly expressed source concept serving as a reference point and the implicit target concept to be accessed belong, has been conceived of in various ways and referred to with a series of terms, such as conceptual domains, dominions, scripts, frames, image schemas, mental spaces, or idealized cognitive models. The notion of idealized cognitive ←15 | 16→models was coined by Lakoff (1987), who defines them as ordered chunks of our knowledge about the world.⁷

According to what has been summarized so far, in the following example (1) the metonymic expression Lemonade (and the concept ‘LEMONADE’ designated by it) serves as a metonymic reference point or source: