Vyvyan Evans: "Semantic representation in LCCM Theory" (2009)

This is a relatively sketchy paper, published in New Directions in Cognitive Linguistics. Its main point is that we should see linguistic representation (discrete grammatical categories) as a crude, schematic system put on top of an ancient system for conceptual representation (qualitative sensibilities).

Various aspects of sentence and word meaning are put in one box or the other depending on whether the aspect is discrete or not. It is thus near-tautological that the "linguistic system" deals with crude, discrete categories, while everything else is relinquished to the dust bin category of fuzzy "conceptual content."

The psychological or "conceptual" part of the story is inspired by Lawrence Barsalou's ideas about mental representations.

Peter Harder: "Meaning as input" (2009)

This paper by Peter Harder argues---building on William Croft and D. Alan Cruse's Cognitive Linguistics (2004)---that sentences should be thought of as little programs that are run in the head of a hearer.

The main example supporting this view of semantics is the case of the definite article. In sentences like

• The UN condemned the test,

the noun phrase the test can be read as an instruction to find a suitable referent (p. 18).

This invites a view of grammar in which the syntactical structure of a sentence is seen as a chronological structure of a program. The sentence The test happened thus receives the following interpretation (p. 23):

• declarative (past tense (HAPPEN (definite (TEST))))

This should be read as a LISP-style program. Harder illustrates the syntax of this instructional language with the following "recipe" (p. 24):

• serve (sprinkle with lemon (grill (add salt (slice (salmon)))))

It should be quite obvious how this is intended to work. Note the parallel to Montague grammar.

The consequence of adopting such a view seems to be that language becomes a narrow communication channel that only allows for the transmission of code, not of meaning. If this is correct, understanding can be compared to a client-side execution of a server-provided script.

Meaning is thus constrained by the resources of the hearer, but the instruction is designed by the speaker. This may be what Harder means by the following quote, although I'm not entirely sure:

Knowing a language, I suggest, essentially consists of knowing the input properties of the forms you choose [the meaning of the keywords in the scripting language?] -- whereas actual outputs [interpretations, meanings?] can never be known for certain in advance [because you cannot control the settings and properties of the client-side computer?]. (p. 16)

The paper was published in New Directions in Cognitive Linguistics (2009), edited by Vyvyan Evans and Stéphanie Pourcel.

Ortony, Vondruska, Foss, and Jones: "Salience, Similies, and the Asymmetry of Similarity" (1985)

A paper from Journal of Memory and Language 24(5) most notable for its long list of good examples of similies (reproduced in the appendix).

Medin and Ortony: "Psychological essentialism" (1989)

This is a contribution by Douglas Medin and Andrew Ortony to the volume Similarity and Analogical Reasoning (1989) edited by Stella Vosniadou and Ortony. Medin and Ortony argue that entities have two distinct sets of features, a shallow and a deep set, that influence similarity judgments in different contexts.

Superficial Attributes and "Essence Slots"
The paper is called "Psychological essentialism" because Medin and Ortony thinks that the context-dependence of similarity judgment can be explained by supposing that ordinary lay folk are metaphysical essentialists. What this means is that the average test subject believes that entities have deep, perhaps unknown, properties in addition to their superficial attributes.

This is supposed to explain why we categorize dolphins with bats and not with sharks. In other cases, like an airplane, a subject's representation of this "essence" make take the form of a theory like "I don't know, but an experts could tell me."

There is something interesting and original about taking a bad philosophical theory and trying to explain it as a psychological phenomenon. However, think their theory hides as much as it shows, as it takes similarity judgment to be divorced from action. I think most of the paradoxical features of similarity judgments (as described by Lawrence Barsalou) would evaporate if we took "being similar" as explained by "treated similarly" rather than the other way around.

A Note on Gender and Style
When I was reading the paper, I was noticing that Medin and Ortony tend to refer to Linda E. Smith by both her first and last name, while they refer to male authors by their last name only. In order to check whether this was actually true, I counted how many times people were mentioned in the text, and whether their first names were mentioned:

Person	+ first name	– first name
Linda B. Smith	4	2
Lawrence W. Barsalou	1	9
Lance J. Rips	1	12
Ryszard Michalski	2	3
Edward E. Smith	2	8
Daniel N. Osherson	2	8
Ludwig Wittgenstein	0	2
John Locke	0	4

The numbers in the table show how often the respective authors occur with their first name spelled out completely. Last names mentioned in references such as "Smith and Medin (1981)" are not counted, since their form is dictated by more rigid style guides. I have not counted names that only occur once (which exlude Elanor Rosch, who is mentioned with first name).

Note that Linda B. Smith is the only woman referred to in the paper. The fact that her first name is mentioned more frequently than, say, Michalski's, may be attributed to the fact that there is another "Smith" frequently referenced in the paper, but note that he is more frequently referred to by his last name only. However, this may partly be due to the fact that his name is disambiguated by the fact that it occurs next to Osherson's as well as more often. A recent mention of a figure does to some extent make the first name obsolete.

Nevertheless, the numbers are quite striking. It would be interesting to do a more thorough investigation of this phenomenon. It would perhaps also be more interesting to investigate whether there is a significant difference in the distance from last mention that warrants reiterating a first name for men and women, respectively.

Srinivas Sankara Narayanan: "KARMA," an addition

I forgot to include this quote in the previous post, but it strikes me as illuminating:

[...] our focus on narrow set of expressions, our impoverished target domain and the inability of our system to extend existing metaphors are important shortcomings. (p. 217)

Srinivas Sankara Narayanan: "KARMA" (1997)

Srini Narayanan's PhD dissertation is often cited as an exemplary accomplishment of cognitive metaphor theory. However, when you read the small print, it seems less convincing. It has the distinct feel of the Georgetown-IBM system from 1953.

Since I have now made my way through the 300-page dissertation, I'll now do my best to give a clear and short presentation of its results. I'll do that in reverse order, starting with its evaluation methodology and moving towards a more detailed description of the program.

Evaluation
Narayanan's program was evaluated on a "corpus" consisting of 9 short snippets of text amounting to a total of 116 words. This evaluation set was, as far as I can understand, specifically constructed so as to look as much like the training set as possible in terms of vocabulary and theme.

For each sentence to be evaluated, the methodology consisted in the following steps:

1. For each sentence, a feature structure was constructed; thus, the sentence was parsed and partially interpreted by hand before entering the reasoning module.
2. Any "difficult" aspects of the sentences were smoothed out; e.g. the phrase perception of was removed from one of the examples to create a flatter semantics (p. 246).
3. Any knowledge that the system lacked in order to understand the sentence was fed into it; e.g., in one case it was given the information that NAFTA affects Mexico (p. 248).
4. The subject matter was, as far as I understand, always explicitly stated to be economic policy so that the system did not have to infer whether it should read motion verbs metaphorically or not.
5. A feature structure was produced by the system, and Narayanan read the contents of this structure and determined whether he thought it had fit his intuitions (as he mostly did).

It should be quite clear that this evaluation procedure is quite liberal. The line between system design and system evaluation is effectively invisible, as is the line between subject and experimenter.

Input knowledge
The system was equipped with a number of materials which were designed explicitly to deal with a training "corpus" of 20 sentences (roughly 750 words). These were:

• A background model of the domain of economic policies with some very strong and very simplified assumptions (pp. 173-75).
• A set of motion models or "execution schemas" that describe the temporal progression of actions like walking and recovering from illness in a highly idealized fashion.
• A set of highly specific maps that told the system how to link various objects, attributes, and events from the execution schemas to the economic objects, attributes, and events. These were often very specific and tailored to the corpus. For instance, there were mappings telling the system to an IncompetentDoctor object onto a Government object, and a step event onto an implement_policy event.

This background knowledge would obviously not be very robust if applied more widely.

The X-schema model
The target domain in Narayanan's thesis, economic policies, is represented fundamentally different from the source domains, movement and health. The target domain is represented as a Bayesian net in order to allow probabilistic reasoning. The source domain is represented as a kind of highly idiosyncratic simulation tool.

These simulation tools are what he calls "x-schemas" or "execution schemas." They are abstract representations of the causal structure of processes like walking or grasping something.

The x-schemas are connceted to a feature structure that keeps track of the state of whoever the agent performing the action is. The agent's state may influence the structure of the process and vice versa. The x-schemas are mostly structured like a little story with a beginning, a middle, and an end, and they are run as simulations when the system processes a sentence.

This simulation is a discrete-step process where the value of various parameters at one time (step size, energy level, distance to goal, etc.) combine to produce the parameter settings in the next time step. All parameter values are non-negative integers. The mechanism that produces the state at time t + 1 from the state at time t is designed by hand (and by intuition).

Examples
Narayanan gives a single completely explicit example of the kind of data structures that the system inputs and outputs. His example builds on the newspaper headline Indian Government Stumbling in Implementing Liberalization Plan.

He manually translates this sentences into the following feature structure representation (p. 208):

( (context Economic Policy)

  (type Liberalization)

  (actor IG)

  (event Stumble)
  (aspect Progressive)

  (ut-type Description) )

Fed this data structure, the programs runs a simulation of a walking event and maps the results to the domain of economic policies using maps like ACTOR IS MOVER. The output is the following (p. 215):

( (context Economic Policy)

  (type Liberalization)

  (actor IG)
  (event Stumble)

  (aspect Progressive)

  (status (ongoing 0 1)
  (suspend 1 1))

  (difficulty

    (t 0 .7)

    (t 1 1))

  (outcome (fail 1 .6))
  (goal

    (free-trade 0 1)

    (dereg 0 1)

    (free-trade 1 1)

    (dereg 1 1)))

It is not entirely clear to me what the first argument means in constructions like (dereg 0 1), and the text doesn't seem to help me out.

Narayanan has previously (pp. 55-58) gone through an example simulation of the WALK schema with a set of explicit values on various parameters such as Dist(walked) and Energy. I will not reproduce the tables here, but the schema can roughly be compared to an loop working according to the following rules:

• If walking is enabled, get ready
• If you're ready, check if you can see
• If you can see, take a step
• If you can't see, check your footing
• If your footing is OK, take a step
• If your footing is not OK, change position
• If you've just taken a step and you're not a your destination, get ready

I should be quite obvious that this is a highly idealized model. Whoever, its experiential implausibility does not constitute a problem in Narayanan's dissertation, probably because of the small size of the validation materials.

The Murphy/Gibbs debate

Spread out over three issues of Cognition in 1996 and 1997, Greg Murphy and Ray Gibbs had a very interesting exchange of opinions about cognitive metaphor theory (Murphy 1996, Gibbs 1996, Murphy 1997).

Murphy argued that a prior-similarity-based account of meaning extension is leaner and more precise than cognitive metaphor theory. Gibbs disagreed.

The Strong View and the Weak View
Murphy (1996) makes a distinction between what he calls the "strong" and the "weak" version of cognitive metaphor theory.

The strong view is that target domains have no inherent structure, and so get all their meaning through metaphors. The weak view is that they do have some "skeleton" of structure (1996: 187), perhaps consisting of something like thematic roles and unspecified relations between actors.

Murphy dismisses the strong view on two accounts.

First, all the false inferences predicted by a general metaphor need to be blocked. This can essentially only be done if the target domain has an autonomous structure, and if there is an authority higher than metaphor that can choose to apply or not apply the metaphors.

Second, suppose we have a chain of conceptualization such as

ARGUMENT <<< WAR <<< A WILD ANIMAL <<< A PERSON <<< A PLANT

There are then four links in which noise, wrong inferences, and memory problems may slip in. This seems computationally and psychologically unfortunate.

"Problems of circularity of evidence"
Murphy also notes that one cannot expect the same pool of data to be both explainer and explained. Non-linguistic evidence is needed (1996: 183).

He notes that some of Boaz Keysar and Bridget Bly's (1995) evidence also seems to suggest that metaphor understanding is a post-hoc, backward-looking construction.

That seems to be how our intuitions work. But Murphy warns us that we should not assume that the products of these intuitions reflect anything about the actual mental process (1996: 184).

"Problem of multiple metaphors"
Murphy goes on to note that some super-metaphorical arbiter must choose between metaphors when we have multiple inconsistent options.

As an example, he mentions LOVE, which is conceptualized as a JOURNEY, an OPPONENT, a UNITY, a HIDDEN OBJECT, a VALUABLE COMMODITY, and INSANITY. These various source domains seem to license different inferences.

In his response to Murphy's article, Gibbs rebuts that "concepts are temporary, independent constructions in working memory created on the spot" (Gibbs 1996: 313).

This is in direct conflict with the claim that the conceptual system exists "in long-term memory" which appears elsewhere in the literature (Lakoff 1998: 51; Lakoff and Johnson 1999: 69).

What counts as literal?
According to Murphy, when Lakoff and Johnson state that physical meanings are the only literal meanings, this is

their theory of concepts applied to language; namely, it says that only simple physical experiences can be directly encoded in linguistic meaning, and nonphysical or abstract relations must be expressed via metaphor. Thus, their claim that Inflation is rising is metaphoric is basically an assumption of their theory, rather than evidence for it. (1996; p. 189)

He concludes that "a number of the 'metaphors' that L&J and others identify may well not metaphors at all" (p. 190). He doesn't seem to have a clear account of how one could test "metaphoricity" himself, though.

Quotes from some reviews of Metaphors We Live By (1980)

From Mitchell Silver's review (1982) in Leonardo:

I believe Metaphors We Live By only articulates and details a form of subjectivism. The subjectivism they reject is a naive, caricatured doctrine. It seems to me that their theses are wholly compatible with the more sophisticated subjectivisms of W. V. O. Quine and Nelson Goodman. (Oddly, they categorize Quine as an objectivist.) (p. 323)

From Barbara M. H. Strang's review (1982) in The Modern Language Review:

[...] one of the issues evaded in the book is the curious one of how we know --- for, epigrams apart, we generally agree --- which is the 'literal' and which the 'metaphorical' meaning.  (p. 135)

From a lively, sensible beginning (less revolutionary that [sic] we had been led to expect) the book moves into a series of essays, capable of throwing light, but drawing vast conclusions from casual evidence. (p. 135)

[...] vast conclusions can be drawn from suggestive, carefully chosen instances. (p. 135)

A golden opportunity has been lost by the resort to inflated and sweeping claims, and to huge leaps from particular instances to vast conclusions (p. 136).

From Robert Greene's review (1981) in Comparative Literature:

It all sounds very Deweyan (p. 1178).

But there seems to be some confusion over which part of the tradition they are actually opposing. They may be closer to Aristotle and Plato than they think. (p. 1178)

He also emphasizes that there often is some uncertainty about whether a certain utterance is a metaphor or not (p. 1175).

From Max Black's review (1981) in The Journal of Aesthetics and Art Criticism:

[...] a reader cannot fail to be provoked and stimulated. (p. 208)

[...] their reiterated psychological or mentalistic emphasis, does no effective work. (p. 209)

[They refer to "experience":] But we are told nothing about what these "experiences" are: referral to the mental simply serves here to mask the absence of a needed explanation. (p. 209)

[With respect to the ARGUMENT IS WAR data:] one might end just as plausibly, with the alternative formulas, AN ARGUMENT IS A DUEL or, VERBAL DISPUTE IS A BATTLE.  (p. 209)

[About "win and lose arguments":] This argument, if it deserves to be called such, is patently circular. (p. 209)

[...] the copious literature on metaphor is almost completely ignored (p. 210).

From John M. Lawler's review (1983) in Language:

When L&J mention people like Whorf, Sapir, Fillmore, Winograd, Wittgenstein, Malinowski, Levi-Strauss, and Ricoeur as intellectual forebears, and then neglect them in the bibliography, they are doing readers a disservice. It would be a significant improvement in future editions to provide an augmented bibliography. (p. 204)

From Michael K. Smith's review (1982) in American Speech:

Within the limits of current American linguistics, their approach is not only novel but perhaps even revolutionary. Considering the broader fields of Western philosophy and language of the entire twentieth century, however, their critiques of both OBJECTIVISM and SUBJECTIVISM are extremely sketchy and do not use to full advantage the more detailed criticisms of these positions which have been made by much of modern philosophy. (pp. 130-31)

Thus, without a systematic strategy for collecting evidence, Lakoff and Johnson seriously undermine the generality of their own arguments. (p. 131)

These two books [by Ullmann and Stern] are not only broader and better researched discussions of metaphor, but they also emphasize one aspect of metaphor that Lakoff and Johnson don't give full scope to: the importance of the human body in many metaphoric transfers. (p. 132)

As a serious scholarly work, then, this book has shortcomings, with the most important a lack of adequate appreciation for prior research (p. 132).

The review also contains references to a large number of very interesting works in American linguistics that precede Lakoff and Johnson's book.

From Wayne C. Booth's review (1983) in Ethics:

It is written in a style so plain as to suggest labels like "ordinaries," or "linguistish." The sentences plod along like those in bad high school textbooks (p. 621).

When they do hazard metaphors [...] they do not in general turn their analysis back upon themselves. Perhaps if they had done so, they would have recognized something about both the relative poverty of their style and the seriousness of their neglect of traditions of literary and rhetorical theory. (p. 621)

[...] the authors seem unaware of just how badly they oversimplify some issues. (p. 621)

He explicitly mention's Dewey's Art As Experience and Ricoeur as references that lack (p. 621). He also blames it for "limited perspective" and "pedestrian [...] style" (p. 621).

Quotes from some reviews of Philosophy in the Flesh and More Than Cool Reason

Two quotes from Joseph Ulric Neisser's review (2001) of Philosophy in the Flesh (1999) in Language:

They [= Lakoff and Johnson] characterize basic metaphors as unconscious and automatic.A fairly rigid structural determinism ensues. (p. 167)

Like latter-day psychoanalysts, L & J spend several hundred pages unmasking the root metaphors which have dominated the minds of unwitting philosophers. (p. 167)

From Michael O'Donovan-Anderson's review of the same book (2000) in Review of Metaphysics:

Although they spend a great deal of time developing an "embodied realism," it is nothing like acceptable in its current formulation. (p. 943)

... their very strange and programmatic chapters on the history of philosophy do not constitute philosophical attention ...(p. 943)

From Richard Taylor's review (1991) of More Than Cool Reason (1989) in The Modern Language Review:

Next time, however, one can only hope that a style somewhat less irritatingly flat and awkward might be used. [...] Schoolbook strategies are even more out of place in so sophisticated an argument; now boys and girls: 'Let us look at the list of basic conceptual metaphors that we have encountered in this chapter.' (p. 653)

From Thomas Leddy's review of the same book (1990) in The Journal of Aesthetics and Art Criticism:

Unfortunately, individual theorists and their specific positions are not discussed. [...] It turns out that in a final chapter titled "More on Traditional Views" the various positions are in fact identified with actual theories of actual authors. [...] Yet in some cases the match is hard to see and in others it is incorrect. (p. 260)

Actually, the similarities between Black's position and Lakoff and Turner's are more striking than the differences. (p. 261)

We are told that death is night draws on the semantically autonomous concept of night which is grounded in what "we experience night to be" (p. 113), i.e., as "dark, cold, foreboding, and so on" (p. 113). Yet although it is difficult to conceive of a night that is not dark it is certainly not the case that nights are always cold and foreboding. The combination of dark, cold and foreboding seems like a fairy-tale, adventure story, or monster movie view of night. In short, although itself without temporal components, it seems to be connected with a type of narrative fiction. (p. 261)

Users of this chapter [ch. 3] should however carefully attend to the authors' cautionary note that this analysis "does not constitute a literary-critical treatment of the poem" and that it is only intended to be "a prerequisite to any such discussion" (p. 159).  (p. 261)

Marina Rakova's debate with Mark Johnson and George Lakoff

In the September 2002 issue of Cognitive Linguistics, Marina Rakova had a short and bloody exchange of ideas with George Lakoff and Mark Johnson.

She criticized their theory for being "philosophically inconsistent" and "contradicted by empirical evidence" (p. 215), and they accused her of "systematic misreadings" (p. 260).

Rakova's paper is not very well-structured, but it does contain some interesting points that I do find worth mentioning:

1. Consistency with developmental psychology

It seems that there may possibly be some measure of evidence that the developmental story implied by cognitive metaphor theory doesn't quite fit our knowledge of actual child development.

She mentions for instance (p. 220) that children are able quite early to use sentences such as There's a butterfly in the garden even though the theory (perhaps) predicts otherwise.

It is not entirely clear to me how much this argument depends on the claim that one needs abstract thought in order to apply a container schema (cf. pp. 232 and 235).

2. Conflicting motivations

Picking up very small things is exceedingly difficult. So why is there no DIFFICULT IS VERY SMALL metaphor, when such post-hoc analyses can motivate the metaphors IMPORTANT IS BIG and DIFFICULT IS HEAVY? 

3. What governs the governor?

If we can have independent knowledge of a target domain, and if we need not use a metaphor automatically, then exactly can we conclude about the role of metaphor?

If metaphors can be switched on and off depending on whether they are useful in the specific context, is there any real content to the claim that we "think metaphorically"?

This is, as far as I understand, the question Rakova raises on the bottom of p. 227.

See also p. 234 and p. 236, where she seems to hint that a bodily experience doesn't just become a crisp, logical concept without some sort of additional enforcement. Lakoff and Johnson mention this on p. 259.

4. Are primitive really universal or just sort-of, so-far universal?

If some conceptual mappings are universal and others are not, and those that are universal are only partly universal, is there any content to the claim that certain concepts are universal?

If there is no theoretical support for the claim, then it should just have the status of "All swans I have seen so far seem to be white," which is not particularly strong or interesting.

In Rakova's words, the theory badly needs to mark off some "boundaries between significant and nonsignificant cultural variations" (p. 229).

By the way, Johnson and Lakoff refer in their response to "mountains of evidence" and "evidence that fills the pages of our discipline to overflowing" (pp. 251 and 261).

Criticisms of cognitive metaphor theory cited by Yu

By and large, Ning Yu's book on metaphors in Chinese is faithful to cognitive metaphor theory. However, he does cite some critics of the theory (as of 1999) and discusses a few of them in chapter 2.8. He rejects them all on the account that they define metaphor differently than Lakoff and Johnson.

The authors that Yu lists as critics of cognitive metaphor theory are:

• Dorothy Holland: "All is metaphor" (1982), a review of Metaphors We Live By
• Earl R. Mac Cormac: A Cognitive Theory of Metaphor (1985)
• Anna Wierzbicka: "Metaphors Linguists Live By" (1986)
• Andrew Ortony: The Cognitive Structure of Emotions (1988), see also these experts
• Jackendoff and Aaron's review of More Than Cool Reason (1991)
• Naomi Quinn: "The cultural basis of metaphor," in James W. Fernandez: Beyond Metaphor (1991)
• Hoyt Alverson: "Metaphor and experience," also in the anthology by Fernandez (1991)
• Bipi Indurkhya: Metaphor and Cognition (1992)
• John M. Kennedy and John Vervaeke: "Metaphor and knowledge atained via the body" (1993)
• Cristina Cacciari and Sam Glucksberg: "Understanding figurative language," in Morton Ann Gernsbacher (ed.): Handbook of Psycholinguistics (1994)
• Gregory Murphy: "On metaphoric representation" (1996) and "Reasons to doubt the present evidence for metaphoric representation" (1997)

Gibbs, Lima, and Francozo: "Metaphor is grounded in embodied experience" (2006)

This is a very interesting paper that reports on an experiment investigating the relation between folk theories of hunger and folk theories of desire. It is written in defense of cognitive metaphor theory but raises a whole array of interesting problems.

Postulates: Embodied primitives and non-embodied compounds
The authors claim that

the poetic value and the communicative expressiveness of metaphoric language partly arises from its roots in people's ordinary, felt sensations of their bodies in action. (p. 1190)

And, in the conclusion, they add that "a significant aspect of metaphoric language is moetivated by embodied experience" (p. 1208).

Following Joe Grady, they also claim that primary metaphors are more likely to be translatable or perhaps universal, since compound metaphors may have different ingredients in different languages:

One implication of our findings is that the ability to translate various linguistic metaphors from one language to another rests on the degree to which these verbal expressions instantiate primary metaphors. (p. 1208)

Method
The authors use a quite interesting experimental paradigm to test their claims with respect to the example metaphor DESIRE IS HUNGER.

Basically, they a group of students to rate a number physical symptoms for relatedness to hunger and another group to rate the same symptoms for relatedness to desire.

This first gives them a list of symptoms seen as typical for hunger in Brazil and in California. Many of these are the same, such as the stomach grumbling or the mouth watering.

The relativity of hunger
But there are also exceptions. Thus, the Brazilian Portuguese speakers generally found that a dry mouth was related to hunger, while being annoyed or depressed isn't. The American English speakers found the opposite.

According to the authors, these differences "reflect something of how cultural experiences shape some of our embodied understandings of hunger," and they speculate a bit about the cultural causes (p. 1204). They repeat this point later:

For instance, the linguistic items She drooled anytime she saw Bob and My stomach aches in anticipation of having sex with Mary were rated as much more acceptable in English than in Portuguese. These differences indicate that the salience of particular, localized bodily experiences of desire as hunger may differ across the two languages. (p. 1207)

Testing for correlation between hunger and desire
After having rated various symptoms with respect to their relation to hunger, the authors are able to see if this relation is mirrored in linguistic items expressing sexual desire, romantic desire, and other desire.

The authors affirm that this is indeed the case, but the details are telling.

For English and Portuguese, the various symptoms are put into two categories, one set strongly correlated with hunger (stomach grumbling) and another only very weakly correlated with hunger (knees swelling). This yields two "weak" sets and two "strong" sets.

The theory is then that this correlation can be rediscovered in the realm of desire. This hypothesis is tested in two ways.

First, some subjects are asked to rate whether certain symptoms are characteristic of desire. For instance, they are asked whether people feel dizzy when they are "deeply in love" (p. 1205). That is the "body questions" experiment.

Second, a number of sentences are produced by the experimenters in which the symptoms are used to talk about desire. For instance, the experimenters ask the subjects to rate the naturalness of the sentence I have a strong headache for you. That is the "language questions" experiment.

Results
In their table of results (p. 1206), the authors have lumped all symptoms in the same block together.

For instance, they have treated all sentences in the cell "English" x "Linguistic" x "Lust" x "Strong" as data points from the same distribution and checked whether that distribution differs significantly from the distribution in the cell "English" x "Linguistic" x "Lust" x "Weak". These contrasts generally do exist.

Critique
Averaging over the symptoms gives a general sense of whether hunger and desire correlates, given the dimensions selected by the experimenters. The authors admit that

our findings on DESIRE IS HUNGER do not imply a complete equivalence between desire and hunger, but only a [sic] strong correlations between them" (p. 1207).

This averaging over all items also smooths out the problem that many sentences are natural in one realm, but not the other.

I would, for instance, like to know how exactly the subjects responded to sentences like My stomach was aching for you. But these interesting differences are blurred because the data is reported without distinctions between various sentences.

It would also have been natural to have a linguistic test for hunger as well as for desire. Consider for instance pairs such as these:

1. ?I'm becoming weak for you. (p. 1205)
2. ?I'm becoming weak for food.

1. ?I have a strong headache for knowledge. (p. 1205)
2. ?I have a strong headache for sugar.

Giving such pairs to different subjects would give a more even experimental paradigm and filter out the effect of the (sometimes quite helpless) prose in the test sentences.


I would also very much like to know if such pairs could indeed produce significant differences. Currently, the paper can't tell me because it only used linguistic stimuli in one part of the study, and doesn't report the responses to individual items.

Sardinha: "Metaphor probabilities in corpora" (2008)

In his contribution to Confronting Metaphor in Use, Tony Sardinha argues that metaphor researchers should care more about the probabilities that a given word will be used metaphorically in a given corpus genre.

To illustrate his ideas, he reports a large number of metaphor probabilities taken from a highly specialized corpus of Brazilian Portuguese.

The article cites three interesting sources that are quite alien to metaphor theory proper:

• The anthology Probabilistic Linguistics (2003), edited by Rens Bod, Jennifer Hay, and Stefanie Jannedy.
• A number of articles from the early 1990s by the British linguist Michael A. K. Halliday, a veteran advocate for corpus linguistics.
• Michael Hoey: "The textual priming of lexis" (2004); this paper is interesting for having a bottom-up, construction-grammar style of explanation.

Boers and Stengers: "Adding sound to the picture" (2008)

This paper by Frank Boers and Hélène Stengers forms a part of the anthology Confronting Metaphor in Use (2008). It makes two points:

1. Different languages apply source domains with different frequencies;
2. the exact shapes of metaphorical idioms are motivated by sound.

Cross-lingual variation in source domain frequencies
The first point should be quite uncontroversial. Boers and Stengers give the example that Spanish has 42 bullfighting idioms, while English has three (p. 69)

However, as they note themselves (p. 69), they are not taking the frequency of the various idioms into account. Instead, they just count the number of dictionary entries.

Boers and Stengers try to patch this deficiency up by performing some corpus searches for the most common idioms in order to use those as an indicator for actual frequency (p. 69-70). This gives a couple of data points with indications of relative frequencies, but not more.

Idiom forms are based on sound
Boers and Stengers' second point is that phrases like it takes two to tango are selected over phrases like it takes two to waltz because of their attractive sound qualities. Further, they predict that English should have more phonetic motivation because of its relatively fixed word order and limited morphology.

Again, these claims are interesting, but their statistical methods for testing them are flawed. In particular, Boers and Stengers do not have any concept of accidental rhyme and alliteration in a language. This essentially means that the differences in "motivation" could be due to the fact that English simple rhymes more by accident.

They are, apparently, aware of this problem (p. 72), because they manually filter out Spanish post-fixes on adjectives and nouns (which would make anything in Spanish seem like it rhymed). They should have developed a methodology that could have done this more generally and quantitatively instead.

Joseph E. Grady: "Theories Are Buildings Revisited" (1998)

This is another paper in which Grady proposes that we should analyze problematic metaphors as results of a combination of several simple, "primary" metaphors.

This time, he applies his method to the "conduit metaphor," the many-faceted metaphor that is alleged to be behind our talk of information content, getting the idea across, etc.

According to his analysis, this compound metaphor is constructed from the following four metaphors:

• CONSTITUENTS ARE CONTENTS
• BECOMING ACCESSIBLE IS EMERGING
• TRANSMISSION IS ENERGY TRANSFER
• ACHIEVING A PURPOSE IS ACQUIRING A DESIRED OBJECT

This analysis can be tested by comparing the metaphor's English elaboration with its elaboration in other languages..

Haser: Metaphor, Metonymy, and Experientialist Philosophy (2005), chs. 2 and 3

Chapter 2 of Verena Haser's book deals with the often ill-defined distinction between metaphor and metonymy. It is not directly relevant to me.

Chapter 3 pecks a little at some of George Lakoff and Mark Johnson's rhetoric. Two "case studies," that is, two point of criticism, are examined. The first (ch. 3.2.1) is the problems with overgeneration, and the second (ch. 3.2.2) is their use of non-existent straw men.

Only the former of these contains arguments directed at the substance of the theory itself (as opposed to its rhetoric).

In this section, Haser notes that ideas can't be sautéed when they evidently can be half-baked (p. 59). Lakoff and Johnson try to edge out of this problem by saying that abstract target domains aren't sharply defined and therefore behave somewhat unpredictably. Haser replies that this concept is ad hoc and ill-defined (p. 60).

She also notes that one of their prime arguments against their rhetorical straw man, the "strong homonymy view," presupposes the fact that metaphors are fueled by conceptual mappings (p. 60).

Criticisms of cognitive metaphor theory cited by Haser

In the "Introductory remarks" of her aggressively polemic book Metaphor, Metonymy, and Experientialist Philosophy (2005), Verena Haser cites a number of relatively recent works that criticizing cognitive metaphor theory.

The list contains the following items:

• Michiel Leezenberg: Contexts of Metaphor (2001)
• Matthew McGlone: "Concepts as metaphors," in Sam Glucksberg: Understanding Figurative Language (2001)
• Dirk Geeraerts: "Cognitive semantics and the history of philosophical epistemology," in Geiger and Rudzka-Ostyn (eds.): Conceptualizations and Mental Processing in Language (1993)
• Olaf Jäkel: "Kant, Blumenberg, Weinrich," in Gibbs and Steen (eds.): Metaphor in Cognitive Linguistics (1999)
• John R. Taylor: "Category extension by metonomy and metaphor," in Dirven and Pörings (eds.): Metaphor and Metonymy in Comparison and Contrast (2002)
• Petra Drewer: Die Kognitive Metapher als Werkzeug des Denkens (2003)
• Marina Rakova: The Extent of the Literal (2003)

Fortunately, I'm already more or less halfway through this list. Surprisingly, the most prominent scholar on her list happens to sit in an office just across the hall from where I am sitting right now.

Kövecses: Metaphor: A Practical Introduction (2002/2010)

This book is a textbook in cognitive metaphor theory, summarizing and surveying most of what has been going on the field since the publication of Metaphors We Live By.

It faithfully reproduces every example and argument ever proposed by that tradition. It has very few original examples, and the quality of the argumentation is often less than one would desire.

Metaphor "governs our thought"
In the preface to the first edition of the book, Kövecses cites a number of conventional metaphors like the ills of society, the road to success, and a branch of the organization.

He then goes on to claim that these aren't dead at all, contrary to appearances:

The "dead metaphor" account misses an important point: namely, that what is deeply entrenched, hardly noticed, and thus effortlessly used is most active in our thought. The metaphors listed above may be highly conventional and effortlessly used, but this does not mean that they have lost their vigor in thought and that they are dead. On the contrary, they are "alive" in the most important sense---they govern our thought: they are "metaphors we live by." (p. xi)

I suppose that by this logic, a word like capital most vigorously evokes the image of a head, since it is even more conventionalized and more hidden than phrases like head of state or warhead.

The evidence for metaphorical thought
Like all cognitive metaphor theorists, Kövecses wants to have his cake and eat it, too: He wants our knowledge of target domains to arise out of metaphors, but metaphors to be constrained by our knowledge of the target domains.

It is quite shocking how little the arguments for these claims have evolved over the past 30 years. I will need to quote him at some length to fully convey this state of affairs.

So, after citing the metaphor LOVE IS A JOURNEY, Kövecses writes (p. 9):

From this discussion, it might seem that the elements in the target domain have been there all along and that people came up with this metaphor because there were preexisting similarities between the elements in the two domains. This is not so. The domain of love did not have these elements before it was structured by the domain of journey. It was the application of the journey domain to the love domain that provided the concept of love with this particular structure or set of elements. In a way, it was the concept of journey that "created" the concept of love.

Note the lack of qualifiers here. All of our knowledge about love arises through metaphorical mappings, it seems. This should be contrasted with his reliance on "objective, preexisting similarity" when he talks about the constraints on metaphor (p. 79).

He goes on:

 To see that this is so, try to do a thought experiment. Try to imagine the goal, choice, difficulty, or progress aspect of love with making use of the journey domain. Can you think of the goal of a love relationship without at the same time thinking of trying to reach a destination at the end of a journey? Can you think of the progress made in a love relationship without at the same time imagining the distance covered in a journey? Can you think of the choices made in a love relationship without thinking of choosing a direction in a journey? The difficulty of doing this shows that the target of love is not structured independently of and prior to the domain of journey.

Another piece of evidence for the view that the target of love is not structured independently of any source domains is the following. In talking about the elements that structure a target domain, it is often difficult to name the elements without recourse to the language of the source. In the present example, we talk about the goals associated with love, but this is just a slightly "disguised" way of talking about destinations given in the source; the word goal has an additional literal or physical use---not just a metaphorical one. In the same way, the word progress also has a literal or physical meaning, and it comes from a word meaning "step, go."  These examples show that many elements of target concepts come from source domains and are not preexisting.

There are thus two main strands of evidence for this theory of the psychology of speech: Introspection and linguistic or etymological evidence.

I think Raymond Gibbs has said what needed to be said about the first one. Verena Haser, Sam Glucksberg, and others have said what needed to be said about the other.