Date: Fri, 18 Mar 2005 08:46:01 +0100 From: Cristiano Broccias Subject: Studies in Linguistic Motivation
EDITORS: Radden, Günter; Panther, Klaus-Uwe TITLE: Studies in Linguistic Motivation SERIES: Cognitive Linguistics Research 28 PUBLISHER: Mouton de Gruyter YEAR: 2004
Cristiano Broccias, Faculty of Modern Languages, University of Genoa (Italy)
OVERVIEW AND COMMENTARY
This volume comprises twelve papers, divided into four sections, which discuss the role of motivation in language. They are preceded by an introduction, "Reflections on motivation",, where Radden and Panther review the notion of motivation (and its relation to its opposite, i.e. arbitrariness). They also highlight its non-deterministic and multifactorial nature. They regard a linguistic unit (or target) as being "motivated if some of its properties are shaped by a linguistic source (form and/or content) and language-independent factors" (p.4). They then discuss the basic semiotic relations underlying motivation as well as various types of language-independent motivation (e.g. ecological motivation, genetic motivation, experiential motivation, perceptual motivation and cognitive motivation). It may be important to stress that by "language-independent motivation" Radden and Panther mean motivation which is not exclusive to the realm of language.
The papers are grouped into five sections, corresponding to different types of motivation. However, the editors stress that their classification is to some extent "arbitrary" (a pun I found very amusing). In what follows I will offer a short summary for each of the contributions as well as my own reactions. I would like to point out from the very start that the book is very interesting and deserves to be used as a reference point for future research on motivation. That is, my evaluation of the volume is positive overall. The fact that in the paper-by-paper analysis below I will insist, especially in some cases, on possible problems with the analyses put forward in the volume simply stems from my desire to see some important issues clarified in future research.
Section I: Ecological motivation
John Taylor, "The ecology of constructions", regards constructions as the basic units of (Cognitive) Grammar and views them as being arranged in networks. The latter point is crucial for his definition of motivation. He claims that "[a] linguistic structure is motivated to the extent that it is related to other structures in the language" (p.57). The reader should therefore acknowledge that Taylor is using a more restricted definition of motivation than Radden and Panther's in that he emphasises the interdependence of phonological, semantic and symbolic structures (i.e. ecological motivation, which counts as an instance of language-independent motivation in Radden and Panther's analysis since networks are not restricted to language). He illustrates (his notion of) motivation by way of the noun "hamburger" (which originated as "hamburg" + "er" and was later reanalysed as "ham" + "burger") and the "Bang goes" constructions (e.g. "Bang goes my weekend!").
Taylor's paper is not only written very clearly but also contributes significantly to the development of the cognitive linguistic enterprise, in particular of Cognitive Grammar. First, in his preliminary remarks, he emphasises that both phonology and semantics have a certain degree of autonomy, a position which is rarely encountered in cognitive linguistic analyses. Second, he stresses the importance of constructions and of distributional facts, thus bridging the (potential) gap between Langacker's Cognitive Grammar, which aims to define grammatical notions semantically rather than distributionally, and Croft's Radical Construction Grammar, which, on the other hand, recognises the primacy of constructions (on this point see Broccias and Hollmann in preparation). Third, by highlighting the systemic nature of grammar, he implicitly sets the agenda for future cognitive linguistic research, which should not ignore that language is a "système où tout se tient". In fact, the volume would have profited even more in terms of its value if ecological motivation had been referred to throughout the volume.
Ad Foolen, "Expressive binomial NPs in Germanic and Romance languages", investigates NPs like "a bear of a man" and "a hell of a job", i.e. complex NPs made up of two nouns (i.e. binomials) which convey a strong expressive force. He claims that both NPs in the binomial are heads. The first NP is the "expressive" head and the second the "referential" head. He also suggests that the construction is both iconically and ecologically motivated. It is iconically motivated because the first position of the attributive (i.e. expressive here) NP (e.g. "bear", "hell") mirrors its salience in the overall construction (i.e. the construction has a strong expressive form). The construction is also ecologically motivated because it can be related to the possessive NP-of-NP construction and the attributive A-N construction.
I found Foolen's analysis very interesting because the pattern he discusses is undoubtedly challenging in terms of its symbolic structure. Also commendable is his insistence on the impossibility of identifying a unique head unambiguously, which is much in the spirit of cognitive linguistics. In a similar vein, he stresses that more than one factor may be involved in the motivation behind this patter, i.e. iconic motivation and systemic motivation. I must admit however that I didn't entirely understand his discussion of ecological motivation. That is, if we go back to Taylor's paper, for example, one can easily see why a certain form arose. "Hamburger" underwent morphological reanalysis because first of all its stress pattern, primary stress on "ham" and secondary stress on "burg", was the same as that of compounds like "cat-lover". In the case of the "Bang goes" construction, the order of the elements, for example, is the same as that found in presentational sentences, with which the "Bang goes" construction shares (some of) its special properties. By contrast, it is not clear to me why binomials, although obviously (synchronically) related to the possessive NP-of-NP construction (because the form is the same) and the attributive A-N construction (because they are functional similar in specifying an attribute of an entity), are motivated by them. That is, the author does not clarify in my opinion what conditions led to the adoption of the NP-of-NP pattern in the first place. One can agree that the use of this pattern may correlate with highlighting "a conceptual distance [...] between the objective individual and the subjective [...] value judgement" (an idea which the author borrows from Campe 1997: 172), but the question remains as to what evolutionary path actually led from the (possessive) NP-of-NP construction to the expressive binomial. Why did speakers select the NP-of-NP form to code the expressive meaning (or the genitive pattern in Latin or the pattern with a motion preposition in Old High German for that matter) if no "of"-relation apparently holds between the two NPs? This question remains an interesting topic for future research.
Section II: Genetic motivation
Bernd Heine, "On genetic motivation in grammar", offers a short summary of his typological studies in grammaticalisation. In particular he discusses the cognitive forces underlying the emergence of the grammatical categories of numerals, indefinite articles and possession. Numerals evolved from our experience with body parts, indefinite articles typically from the numeral "one" and possession can be related to various source schemas.
The importance of this paper lies in Heine's contention that a claim like "I cannot see any motivation, hence, there is no motivation" (see p.118) should be rejected if serious research is carried out. In fact, he underlines that his previous work in various areas has succeeded precisely in showing that motivation can be found. It should be noted however that Heine's article is a summary of previous research (he also prefers to refer the reader to his other publications for specific examples, see p.107). This may probably be due to the editors' desire to primarily offer a book which can be used as an introduction to the issue of motivation within the cognitive linguistic paradigm.
Christian Koops, "Emergent aspect constructions in Present-Day English", analyses three types of construction in Present-Day English which can be associated with progressive meaning in other languages, i.e. locative constructions (e.g. "I was in the middle of getting my hair cut"), posture verb constructions (e.g. "How could you stand there and watch them beat that guy?") and motion verb constructions (e.g. "You can't go around testing everybody for everything"). The English constructions all evoke imperfectivity (they are progressive, durative and repetitive respectively). They are genetically motivated in that their meaning is shown to follow from the lexical meanings of their source notions.
I think Koops' paper is one of the best in this collection because of his detailed examination of various constructions and the emphasis on their relatedness (through the notion of imperfectivity). I have only two minor observations to make about it. Still, my first point may be of some importance if the book is used as an introduction for non-expert readers.
Koops claims that progressive constructions "are typically restricted to dynamic events and incompatible with states" (pp.123-4). Although Koops' statement is hedged by "typically" (see also p.134 where he recognises that stative "sit" and "stand" are compatible with the progressive), his claim may be symptomatic of the fact that "[i]t is sometimes supposed that the progressive aspect occurs only with dynamic verbs describing activities or events. However, the progressive can also be used with verbs that describe a static situation. In this case, the progressive expresses the meaning of a temporary state that exists for a period of time [...] Some of the most common verbs occurring with progressive aspect are of this type [i.e. stative, CB]" (Biber et al. 1999: 471). That is, the progressive in English seems to be able to evoke both "temporal expansion" (i.e. imperfectivity) and "temporal transience" although either aspect may be highlighted through the choice of a particular verb phrase. Indeed, there may be no reason to view the use of progressive aspect with stative verbs as exceptional (see Williams 2002; on the dangers of a unitary view of progressive aspect and the usefulness of its characterisation in terms of both expansion and transience for other grammatical phenomena see also Broccias 2005).
The second observation is: is the author sure that the constructions he deals with are emergent constructions in Present-Day English? What about past stages of the language? Of course, I fully understand that the author couldn't deal with this in a (necessarily) short paper. It will be interesting to see what diachronic data can tell us about these constructions.
Section III: Experiential motivation
Vyvyan Evans and Andrea Tyler, "Spatial experience, lexical structure and motivation: The case of "in"") argue that the various senses of "in" can be accounted for in a principled way by regarding them as extensions from a "proto-scene", which describes "containment". Such extensions are motivated by our experience with different aspects of containment. They distinguish fifteen senses, which they arrange in a radial network.
The important point made by Evans And Tyler is that sense extension can be shown to have an experiential basis. Although I fully recognise the importance of this claim, I have three main reservations concerning Evans and Tyler's paper.
The first reservation is psychological. The authors are fully aware that "not all senses associated with a particular phonological form may be recognised by a language user as being synchronically related" (note 4 on p.165; an almost identical note expressing the same warning is found in their 2001 paper (note 21, p. 744)). This immediately raises the question of what their network in Figure 3 is intended to represent, i.e. what is its psychological status? This question also bears on the nature of the proto-scene. The proto-scene is defined as a "highly abstract representation" (p.166). But then one must first show that such highly abstract representations do have a significant impact on how we manipulate language or, to put it differently, that they have a significant degree of activation in the conceptualiser's mind. In fact, research such as Boas's (2003) (although carried out in the area of resultative constructions) seems to suggest that extensions originate from very concrete uses rather than highly abstract representations. It may also be relevant to remember that Langacker has repeatedly expressed reservations about the degree of activation of high-level schemas in the conceptualiser's mind (see e.g. Langacker 1999: 118).
The question concerning the psychological status of the proposed network leads me to the second reservation, which is empirical. I think that the authors' framework and analyses would profit greatly not only from psychological experiments (the relevance and importance of which is admittedly duly remarked upon in the note mentioned above) but also from corpus evidence (both synchronic, see below on the "rub in" example, and diachronic, see below on the Means Meaning). Although their line of reasoning (in motivating sense extensions) sounds in general convincing (but see also below), empirical evidence, if available, is a sine qua non for safer analyses (in fact the authors themselves point this out in their 2003 volume, see e.g. p.236). Contrary to what the authors seem to claim in note 4 in this paper and in note 21 of their 2001 paper, (at least some) corpus evidence can already be made available. Gries (2004), for example, has shown that cognitive linguistics can benefit from corpus linguistic analyses of word senses in that semantic networks can be investigated by using corpus evidence. Further, the proposed paths of extension deserve to be checked against both diachronic evidence and cross- linguistic evidence. Of course, I am not suggesting that the authors should have done this in their paper, given obvious space limitations. Rather, I am simply suggesting that some references to this, at least as a topic for future research, could have been made.
Further, some analyses may be objected to. For example, the authors claim that the Means Sense of "in", as in "She wrote in ink" and "He spoke in Italian", is motivated on the basis of "the tight correlation in experience between an activity and the means of accomplishing the activity" (p.178). In more detail, they say that "[t]his has been possible precisely because "in" had an antecedent Activity Sense associated with it [i.e. the one exhibited in sentences like "He's in the governor's office", meaning "He works for the governor", CB]" (p.178). One may want not agree with this analysis or at least object to the examples meant to illustrate it. The first problem has to do with the causal relation envisaged by the authors between the Activity Sense and the Means Sense (see the use of "because" in the sentence above but note, however, that such a causal link is not captured in Figure 3 on page 173). Italian, for example, lacks the Activity Sense, as this sense is illustrated through the authors' examples, for the preposition "in" (the form of the preposition is the same as in English) but still has at least some Means Sense uses. English "in" in "She wrote in ink" is rendered with the Italian preposition "con" (English "with"). "In" in "He spoke in Italian" is, on the other hand, also rendered with the preposition "in" in Italian (i.e. "Parlò in italiano"). Of course, one could object that the prepositional system is not the same in the two languages and hence the paths of extension for Italian "in" and English "in" differ. If, however, Evans and Tyler's explanation for the use of "in" in "He spoke in Italian" is language independent, then the Italian rendering may cast doubt on their analysis.
In either case, this simple example shows that a purely speculative analysis is not sufficient. What's more, one may also wonder whether the sentence "He spoke in Italian" actually involves a Means Sense. Why can't this use of "in" be based on an alternative explanation (or a combination of more than one kind of motivation)? For instance, one might view "to speak in language X" as "to use words that are found in language X", i.e. "in" has the (prototypical) container function in that we view words as objects in a language/container. From a historical point of view, I would also like to observe that the Oxford English Dictionary (OED) does not seem to mention the Activity Sense meaning, at least as Evans and Tyler seem to intend it on the basis of the examples they provide (see also below). By contrast, the sentences used by the authors to illustrate the Means Sense (i.e. (17a), "She wrote in ink", and (17b), "He spoke in Italian") are apparently related to two types in the OED (see OED senses 13 and/or 14 for the former (and observe that sense 14 is attested for the first time in1663) and 12c for the latter). Note also that examples like (17b), "He spoke in Italian", date back to at least c900 according to the OED. So it may seem strange that the Activity Sense from which Evans and Tyler claim the Means Sense is derived is not recognised in the OED. In fact, on the basis of the OED data (i.e. OED sense 7), it may be the case that Evans and Tyler's examples like "She's in medicine" are later (metonymic) developments from "membership" examples like "to be in a company/college/association/party, to be in the army/navy", which indeed are contemporaneous with the Means Sense (the first example involves Old English "here", i.e. "army").
One more example whose analysis I find debatable is (25b), "Angela rubbed in the lotion". The authors claim that in this example "the lotion is not entering the skin, only to be free to leave again. The skin is not being conceived as an entity with interior space. Rather, [... "in" can be used in this example because] "in" derives a Disappearance Sense which can come to be used in contexts unrelated to the original context which motivated this sense in the first place" (p.184). I doubt that the skin cannot be conceptualised as being three-dimensional (cf. "Ultracare 3 is quickly absorbed into the skin and forms a protective grease-free barrier" from the Longman Dictionary of Contemporary English, CD-version; this example seems to show that lotions are conceptualised as both moving into a three- dimensional location and creating a surface layer). Further, the Oxford Dictionary of Phrasal Verbs glosses "rub in(/into)" as "force (something) into (a material) by rubbing it over the surface of the material" and gives as possible objects "oil", "polish", "cream", "ointment", "linament". This clearly shows that the Oxford Dictionary of Phrasal Verbs conceptualises "rub in" as involving motion into a three-dimensional location.
I'd like to conclude this rather lengthy discussion of Evans and Tyler's paper with one more (ecological) observation, which is linked to my second reservation (being empirical in nature) and which I hope the authors will take into due consideration in their future research. It may be the case that certain uses of "in" evolved because a "niche" (to use Taylor's word) was available for them. This is the case with the many instances in which "in" is used as either an intransitive preposition or in compounds (see Evans and Tyler's examples (20)-(25) and (31)). The fact that e.g. "Angela rubbed in the lotion" is possible in English but does not have a word-by-word translation into Italian can be related to the fact that "change constructions" (i.e. what are usually called resultative sentences, see Broccias 2003) are very much restricted in Italian (as well as in other Romance languages). Further, it should also be stressed that any account which purports to trace the paths of extension of the preposition "in" should take into account certain very important historical facts of an ecological nature. In particular, it is worth remembering that the preposition "in" was seldom used in Old English (which preferred "on"). That is, the evolution of "in" should not perhaps be studied independently of the preposition "on" which it came to replace in many cases. This point, and the preceding discussion, show that cross- linguistic data, synchronic data and diachronic data are all essential ingredients to investigate meaning extension even within an experientially based account and that their use should not be delayed.
John Newman, "Motivating the uses of basic verbs: Linguistic and extralinguistic considerations", shows how basic verbs (e.g. "eat", "give", the state verbs "stand", "sit", "lie") motivate various grammatical facts (serialization, tense and aspect markers, noun classifiers, case marking) in various languages.
His paper is very important because it does not only report on past research but also advances new suggestions as to how posture verbs may interact with agent-patient distinctions in various languages. Further, it fits well into the volume because Newman takes a very balanced view of motivation: he explicitly combines experiential motivation with ecological motivation (see for example p.194 and p.201). Only if ecological motivation enters the equation can we motivate the origin of grammatical structures satisfactorily. I also liked very much his being explicit about the implications concerning the interaction between experiential and ecological motivation. His final remark on this is worth quoting almost in toto: "taking certain structures as given [...] begs the question of how these linguistic structures came to be there in the first place. Ultimately, the linguistic structures assumed in the discussion here need to be accounted for and also motivated by other linguistic [i.e. ecological, CB] and extralinguistic considerations. In this way, one is led to a kind of infinite regression into the very foundations of language" (p.214).
Section IV: Cognitive motivation
Teenie Matlock, "The conceptual motivation of fictive motion", discusses fictive motion sentences (e.g. "A trail goes through the desert"), where a motion verb is used to describe a static scene. She illustrates the grammatical properties of the construction and argues that they are motivated by the fact that mentally simulated motion is part of fictive motion sentence processing.
Matlock's contribution mainly reports on her past and on-going research. Still, it is very welcome as an introduction to fictive motion sentences, especially because of its focus on processing. Matlock also points out that much research still needs to be done in the area of fictive motion constructions. In particular, I found interesting her suggestion (based on previous work by Langacker) that the acceptability of some sentences may depend on Langacker's sequential scanning. She claims that the contrast between "??The cell phone goes from the cup to the book." (where the phone is on the desk and we imagine that the cell phone is unusually long) and "The cell phone goes from the cup to the book" (phone in ad on billboard) may be due to absence vs. presence of sequential scanning involving the subject NP. She argues that a cell phone is not scanned sequentially (even if it is unusually long) because "a coherent whole can be obtained with just one glance" (p.228). Of course, this is just a suggestion and it is open to future research. But I would like to point out that the nature of scanning is a very thorny issue (see for example Broccias and Hollmann in preparation). Further, if I understand the intended difference between the two sentences correctly, it might be that such a difference does not have to do with scanning per se but rather with the relative dimensions of the objects involved (i.e. perspective). That is, in the phone-on-desk case, the phone is still (much shorter) than the table and the table is thought of as being of ordinary dimensions. In the ad-on-billboard case, the phone (as well as the table) is much larger than usual (as compared for example with our body). Note that this explanation is not dissimilar from the one Matlock herself offers for another example, namely her (6b), "??The small, perfectly round hot tub goes along the back fence" (which is ok if an appropriate perspective is chosen, see p.228).
To put it differently, the problem with the sentence discussed here is that there might be a clash between the verb, which prototypically evokes a non-negligible extension for the entity of which it is predicated, and the entity chosen as its subject. This is most visible in another pair of examples which Matlock explains on the basis of sequential scanning, namely "?The sidewalk runs from here to there." (five feet long) vs. "The sidewalk runs from here to there" (500 feet long). If we assume that the prototypical translational use of "run" requires a relatively long distance to be run, than the choice of a five-foot sidewalk as a subject clashes with this requirement (it sounds odd in normal circumstances to say that a person, for example, ran five feet). Further, note that the very choice of sidewalk to refer to something which is five feet long may be questionable anyway. Be that as it may, I think that Matlock's research is of crucial importance to the development of cognitive linguistic analyses and one cannot praise enough her experimental approach.
Anatol Stefanowitsch and Ada Rohde, "The goal bias in the encoding of motion events", investigate two hypotheses concerning the apparently less restricted distribution of goal-PPs as compared to other path-PPs, namely the (psychological) salience hypothesis and the complete-conceptualisation hypothesis. The former motivates the goal bias in terms of our greater interest in the goal of actions than in sources. The latter claims that goal PPs are less restricted because they are more informative. Using corpus evidence, the authors show that the picture is more complex than usually assumed. Although goal PPs seem to be preferred in general, the existence of "exceptions" is due to the nature of the verbs employed (e.g. the manner of motion verb stroll combines primarily with trajectory PPs). Stefanowitsch and Rohde also argue that the complete-conceptualisation hypothesis has more explanatory power than the salience hypothesis. They recognise however that the two may not be mutually exclusive.
This paper is also very good. It demonstrates the importance of corpus research for studies of motivation in a very clear manner. I have only a relatively minor observation concerning the authors' claim that "[i]t is simply not the case that every motion event is conceptualized as having a source, a trajectory, and a goal [...] There is nothing to stop us from construing a motion event as having only a source, only a trajectory, or only a goal. Verbs like "cruise", "stroll", and "escape" impose just this type of construal" (p.264). I fear that this statement may confound two issues. On the one hand, I agree that not every event has a source- trajectory-goal structure. This is so because translational motion is not the only possible type of motion, of course (consider circular motion, for instance). For example, the author's sentence "They were cruising up and down Main Street" can be regarded as an instance of an oscillatory type of motion. On the other hand, I find doubtful the claim that certain translational events have, for example, only sources (by contrast, one could of course argue that sources are not conceptualised in circular motion). There is no clear sense in which one can use the notion "source" if a target is not activated at some level of saliency. The fact that the target cannot be pinned down exactly is another matter and in any case it remains to be shown that this means that the target is not conceptualised. At worst, the target is taken to be the complement of the original location. If I went out, the target is my not being in any longer. What seems to me (at least intuitively) to be conceptualised perhaps optionally is only the trajectory (if I say "I went out into the back garden", I agree that it may be difficult to identity the trajectory here. As soon as I have walked through the door, I'm in the garden. Still, it is interesting to observe that a trajectory preposition is used with reference to the door, i.e. "through". It is as if the conceptualisation of the trajectory were reduced to a minimum; hence, even in this case, it may be argued that some residual notion of trajectory is activated after all).
In sum, the conceptualisation of events in terms of schemas other than the source-trajectory-goal schema is in principle independent of the authors' observation that sources, for example, are not conceptualised in some (translational) events. In translational motion cases, the claim that sources can be conceptualised without reference to targets is rather a strong one and should be investigated carefully. A source by definition implies an inner and an outer (or goal) space. Similarly, the notion of goal is complementary to that of source. In either case, the fact that the complement of what is expressed in the syntax is not profiled does not necessarily mean that the complement is not activated.
Gerhard van Huyssteen, "Motivating the composition of Afrikaans reduplication: A cognitive grammar analysis") motivates the existence of Afrikaans grammatical and onomatopoeic reduplications (e.g. "plek-plek", lit. "place-place", i.e. "in some places", and "heop-hoep" to refer to the bird scientifically known as "upupa africana") through various metonymies including MORE OF FORM FOR MORE OF CONTENT and PRODUCT FOR PRODUCER. He also implements his analysis using the descriptive apparatus of Cognitive Grammar.
Huyssteen's paper is interesting even if, as a reader of a volume on motivation, I would perhaps have liked less technical discussion on how his analysis can be implemented within the framework of Cognitive Grammar. Still, it must be emphasised that the author does not only refer to metonymic motivation but also hints at ecological motivation as contributing to the existence of reduplicated forms. On page 289, the author cogently remarks that the use of reduplication is also found in other areas, e.g. to code aspect. Such a balanced view of motivation should be much appreciated, in my opinion.
Francisco José Ruiz de Mendoza Ibáñez and Olga Isabel Díez Velasco, "Metonymic motivation in anaphoric reference", study how to account in a principled way for cases where an anaphoric pronoun is used to refer to a metonymic antecedent (e.g. "The ham sandwich is waiting for his check and he/*it is getting restless."). After reviewing their analysis of what a metonymy is - they recognise two types, source-in- target metonymy (e.g. "He's a real brain") and target-in-source metonymy (e.g. "Chrysler has laid off a hundred workers") - they propose a general constraint on metonymic anaphora and three principles which are graded with respect to each other and interact in a way that captures the data considered.
Ruiz de Mendoza Ibáñez and Díez Velasco have written a very thought- provoking paper, which will have to be taken into serious consideration in subsequent analyses within cognitive linguistics. Still, I must point out that I have various reservations about the nature and scope of their analysis, which I hope the authors will be able to clarify in future work.
First is the issue of empirical evidence. The authors base their analysis on very few invented examples. Although it may be very difficult to find natural examples of anaphoric reference to metonymic antecedents, the risk of arriving at empirically debatable conclusions is high. Even within their paper, apparently similar examples are judged differently. For instance, (12b) "*The ham sandwich is waiting for his check and it is getting restless" is starred but (16b) "?The mushroom omelet left without paying its bill. It jumped into a taxi" (from Stirling 1996) only receives a question mark. But both are explained on the basis of the same principle (the Domain Availability Principle or DAP, see p.307; but see also below on this contrast). Still, the authors set up a complex system of principles precisely to explain sentences which receive question marks vs. stars (e.g. (17a), "?The mushroom omelet left without paying. It was inedible" vs. (17b) "*The mushroom omelet left without paying. It was inedible"). That is, not only are the data not authentic but the acceptability judgements could also be objected to. It would perhaps have been useful to include examples rated by a sufficiently large number of native speakers.
The second observation is: admitting that Ruiz de Mendoza Ibáñez and Díez Velasco's principles exist, what is their place in grammar (I use "grammar" in the sense of Langacker's, i.e. as a structured inventory of conventional linguistic units, see also the quotation from van Hoek 1997 below), i.e. how are they represented in the conceptualiser's mind? Do they have an independent existence of the cognitive principles on which they are hypothesised to be based? Are they relevant to both production and understanding? Does their ordering differ from speaker to speaker? Of course, the authors could not possibly have answered all these questions in their paper but I believe that questions like these could have been mentioned at least in passing since they are not trivial at all.
The status of the proposed principles seems to me to be reminiscent of that of principles postulated in generative grammar (it is also interesting that the authors on p.313 do actually use some sort of generative jargon, either voluntarily or not: "Finally, the CMA is not strictly a principle but a filter for attempts to generate metonymies that would cancel out a metonymy which is already active in the antecedent.") Related to this is the fact that to postulate such a complex system in order to motivate very few examples may in itself be suspicious (see also the fifth point below). But why postulate very specific (i.e. metonymy- specific) principles and not rely on more general cognitive notions?
The third point concerns the grading of the principles with respect to each other. The authors' line of reasoning risks to be circular unless more examples are considered. For instance, on page 312 they write that "[w]hen the metonymy is in the antecedent as in (19), we have a situation in which the DAP needs to apply first". In other words, they seem to apply the postulated principles to their data in an order dictated by the relative strength of their principles but the relative strength of their principles is derived from the very data the principles purport to capture. Further, the question of how this grading is represented in our mind remains (see the previous point, especially in connection to possible variation among speakers).
The fourth observation has to do with whether these principles are universal or specific to English. Consider for example what happens in Italian with (16a) "The mushroom omelet left without paying its bill. He jumped into a taxi" and (16b) "?The mushroom omelet left without paying its bill. It jumped into a taxi". Italian usually omits subject pronouns so we can't use them to track reference. However, Italian (at least in its Northern varieties) does not use the past simple to refer to past events but a form corresponding to the English present perfect (i.e. Italian "passato prossimo"). Importantly, the auxiliary employed with verbs of motion is "essere" (i.e. English "to be") and the participial form behaves like an adjective in this case in the sense that it agrees in gender with the subject (both the French loanword "omelette" and its Italian equivalent "frittata" are feminine in Italian). Interestingly, the only translation I find acceptable is "La frittata ai funghi se n'è andata senza pagare e si è infilata in un taxi" (observe that I added the conjunction "e", "and", to render the whole expression more natural. An alternative could have been to use two separate clauses connected by the temporal conjunction "poi", "then". Without either "e" or "poi", I wouldn't accept the sentence). That is, the participial form "infilata" (colloquial for "to rush/jump") is feminine, not masculine. I find the masculine version, in which the participial agrees with the intended referent, not acceptable. In other words, the Italian pattern is the opposite of the English one. This (possibly) shows that if we accept Ruiz de Mendoza Ibáñez and Díez Velasco's principles either we would have to reorder them in other languages or have to think of them as interacting with other "constraints" so that the picture becomes even more complex. I'm not denying that some sort of interaction may actually take place; rather, I'm objecting to motivating the acceptability judgements reported in this paper in terms of very well-defined, i.e. specific, metonymic principles.
My fifth concerns has to do with other possible ways of capturing the examples in (16)-(21) without invoking the authors' complex machinery, i.e. a set of specific principles. What I have in mind as a cognitive analysis of metonymic anaphora would be an investigation along the lines of Hoek's (1997) study of non-metonymic anaphora:
"In this view [i.e. using van Hoek's model of conceptual reference points, CB], the anaphora constraints are not distinct principles which must be independently listed in the grammar; rather they emerge from the nature of linguistic semantic organization in general and nominal semantics in particular. [...] The theoretical machinery underlying grammaticality judgments involves the interaction of schemas, that is, constructional templates, which are entrenched to varying degrees." (van Hoek 1997: 218)
Just for the sake of the argument, let's suppose that the judgements in (16)-(21) are indeed those of the majority of native speakers (but remember that there is a contrast in grammatical judgement between (16b), which has a question mark, and (12b), which has an asterisk). I will now sketch out an explanation for (16)-(17) which does not require the postulation of any specifically metonymic principles but rather relies on general cognitive principles.
The fact that the example in (16a), "The mushroom omelet left without paying its bill. He jumped into a taxi", is better than (16b), "?The mushroom omelet left without paying its bill. It jumped into a taxi" may be simply due to the fact that "he" is an unmarked option for reference with respect to "it". In other words, one naturally expects the metonymic link to decay more quickly than its target (see also Panther and Thornburg's 2002 view of metonymy as being contingent). So we are simply using an unmarked anaphora in (16a), whereas (16b) requires us to keep the metonymic link active for longer (although the similarity of the two clauses, both referring to motion, may contribute to make the whole expression partly acceptable. This might also motivate why (12b), "*The ham sandwich is waiting for his check and it is getting restless", is worse, at least if we agree on this acceptability judgement. That is, in (12b) we don't have the same frame in both clauses: the first is about waiting, the second is about the psychological consequences of waiting). Admittedly, however, the Italian data would need a different explanation. My point here is simply to try to show that without postulating any principles like Ruiz de Mendoza Ibáñez and Díez Velasco's one can come up with a different analysis accounting for the same data.
As to (17a), "?The mushroom omelet left without paying. It was inedible", vs. (17b), "*The mushroom omelet left without paying. He was inedible", we can motivate the deviance of the former by observing that the sentence evokes two different frames (as the authors themselves duly remark). The first clause obviously targets a person, not the food, so "it" in the second clause requires us to do more work than a sentence starting with "he", cf. (16a) (incidentally it might be interesting to study what happens if one continued with "He found it inedible" rather than "It was inedible"). (17b) is out simply because it is very difficult to see the reason why one would want here to predicate a property of an object through the person that has control over (i.e. eats) it. Whereas the first clause conveys some sort of expressive meaning and/or allows the act of reference to take place more quickly (cf. "the person who ate the mushroom omelet"), there is no obvious advantage either in terms of either expressiveness or economy or clarity in choosing "he" over "it" in the second clause.
In sum, I hope to have shown that by relying on very general principles like, for example, "markedness", "economy" and "expressiveness" one could provide an alternative view of (at least some of) the facts discussed by Ruiz de Mendoza Ibáñez and Díez Velasco. I suspect that that these general principles interact non-deterministically every time we are confronted with sentences such as those discussed by the authors and do not actually result in an ordered set of specific (i.e. metonymy-related) principles. But before doing that, it is necessary to (at least) (a) collect natural examples and ask a statistically significant number of native speakers to judge the sentences on which the authors' theory rests and (b) take into consideration what the functional (i.e. communicative) motivation is for using a marked option.
Rita Brdar-Szabó and Mario Brdar, "Predicative adjectives and grammatical- relational polysemy: The role of metonymic processes in motivating cross- linguistic differences", consider various English constructions which take predicative adjectives (e.g. "London was foggy today", "I am hot", "I was firm of purpose", "One should be as clear as possible about historical facts", "The editor is certain to reject it") and observe that they do in general correspond to predicative adjective structures in Croatian, German and Hungarian. The authors claim that English, unlike the other three languages, relies heavily on metonymic processes in structuring its clauses (i.e. English exhibits more "grammatical polysemy" than the other three languages).
Although the title makes reference to metonymic processes only, at the end of their paper Brdar-Szabó and Brdar relate the greater recourse to metonymic processes in English (as opposed to the other three languages considered) to the lack of a flexible word order in English (see pp.350- 351). The recognition of the importance of linguistic considerations (i.e. word order), alongside metonymic ones, is welcome because, by focussing on both aspects, the authors strengthen their analysis.
Anotonio Barcelona, "Metonymy behind grammar: The motivation of the seemingly "irregular" grammatical behavior of English paragon names") concludes the volume investigating the conceptual operations underlying the use of paragon names (e.g. "That young man is a real Shakespeare"). He claims that two metonyms are involved, namely CHARACTERISTIC PROPERTY OF AN INDIVIDUAL FOR THE INDIVIDUAL and IDEAL MEMBER FOR THE CLASS.
The novelty in Barcelona's work resides in his use of two metonymies rather than one to elucidate the use of paragons and in his viewing the CHARACTERISTIC PROPERTY OF AN INDIVIDUAL FOR THE INDIVIDUAL metonym as purely conceptual or prelinguistic (i.e. p.369). That is, we access our stereotypical model of Shakespeare (the target) through a source where a link between SHAKESPEARE and (HAVING) IMMENSE LITERARY TALENT is created. Crucially, however, this metonymy does not show up independently in the language. Barcelona is aware that his proposal may appear controversial and repeatedly insists on it (see for example page 367). Only future research, I think, will be able to tell us if his stimulating analysis is on the right track.
Since I have commented extensively on most of the papers in the previous section, I will now limit myself to some general remarks. All in all, the book under review is an important contribution to the study of motivation from a cognitive linguistic perspective. It is commendable both as an introduction to various types of motivation currently being investigated - Radden and Panther's introduction is also very clear and informative in this respect - and as a starting point for future developments.. As I have pointed out above, there are some excellent papers (e.g. Taylor, Koops, Newman, Matlock, Stefanowitsch and Rohde) and a great variety of challenging ideas have been put forward. Taylor's contribution in particular stands out because it points to a crucial requirement for future cognitive analyses. It is not enough to say that structure X in language Y is motivated, for example, experientially if we do not recognise first that a "niche" for structure X is available in language Y. This point is also mentioned explicitly by Newman, who is aware of its epistemological implications. I think that it is one of the greatest merits of this volume to have brought this point to the fore within cognitive linguistics. Finally, I hope that my observations, especially in the case of the papers by Evans and Tyler and by Ruiz de Mendoza Ibáñez and Díez Velasco, may be of some value to clarify a few issues that not all cognitive linguistics- oriented scholars may agree with. To be sure, this book paves the way to future cognitive analyses which will take into serious consideration various strands of motivation, from ecological motivation to cognitive motivation.
Biber, Douglas, Stig Johansson, Geoffrey Leech, Susan Conrad, and Edward Finegan. 1999. Longman Grammar of Spoken and Written English. Harlow: Longman.
Boas, Hans. 2003. A Constructional Approach to Resultatives. Stanford, CA: CSLI Publications.
Broccias, Cristiano. 2003. The English Change Network. Forcing Changes into Schemas. Berlin: Mouton de Gruyter (Cognitive Linguistics Research 22).
Broccias, Cristiano. 2005. The construal of simultaneity in English with special reference to as-clauses. Unpublished Manuscript, University of Genoa.
Broccias, Cristiano and Willem Hollmann. (in preparation). Do we need summary and sequential scanning in (Cognitive) grammar?
Bullon, Stephen et al. 2003. Longman Dictionary of Contemporary English (with CD-Rom). Harlow: Longman.
Burchfield, R.W. et al. 2002. Oxford English Dictionary (CD-version). Oxford: Oxford University Press.
Campe, Petra. 1997. Genitives and "von"-datives in German: A case of "free" variation? In Verspoor, M. et al. (eds.). Lexical and Syntactical Constructions and the Construction of Meaning. Amsterdam: Benjamins. 165-185.
Cowie, Anthony and Ronald Mackin. 1993. Oxford Dictionary of Phrasal Verbs. Oxford: Oxford University Press.
Gries, Stefan. 2004. Corpus-based methods and cognitive semantics: The many senses of "to run". Unpublished Manuscript, University of Southern Denmark.
Langacker, Ronald. 1999. Grammar and Conceptualization. Berlin: Mouton de Gruyter (Cognitive Linguistics Research 14).
Panther, Klaus-Uwe and Linda Thornburg. 2002. The role of metaphor and metonymy in English "-er" nominals. In Dirven, René and Ralf Pörings (eds.). Metaphor and Metonymy in Comparison and Contrast. Berlin: Mouton de Gruyter (Cognitive Linguistics Research 20). 279-319.
Stirling, Lesley. 1996. Metonymy and anaphora. Belgian Journal of Linguistics 10: 69-88.
Tyler, Andrea and Vyvyan Evans. 2001. Reconsidering prepositional polysemy networks: The case of "over". Language 77: 724-765.
Tyler, Andrea and Vyvyan Evans. 2003. The Semantics of English Prepositions. Spatial scenes, embodied meaning and cognition. Cambridge: Cambridge University Press.
van Hoek, Karen. 1997. Anaphora and Conceptual Structure. Chicago: University of Chicago Press.
Williams, Christopher. 2002. Non-Progressive and Progressive Aspect in English. Fasano: Schena editore.
ABOUT THE REVIEWER:
ABOUT THE REVIEWER
Cristiano Broccias is a Research Fellow in English Language and Linguistics at the Faculty of Modern Languages of the University of Genoa (Italy). His main interests lie in the description and cognitive linguistic analysis of English grammar, both synchronic and diachronic. His publications include a monograph on English change constructions: "The English Change Network. Forcing Changes into Schemas", Mouton de Gruyter (Cognitive Linguistics Research 22), 2003.