Griceian ornithology

Fragments from Mitchell/Bar-On:

"Avian Calls as a Case Study in Expressive Behavior"

"In a survey of a half-century of ethological research, Marler (2004) considers various forms of bird calls (as distinct from bird songs, which tends to be longer and more acoustically complex). In the chaffinch, calls have the following functions: courtship, aggression, predator alarm, announcement and exchange of food, distress, and group proximity maintenance. (There may also be a ‗regenruf‖, predicting rain.) In the domestic chicken we find a repertoire of about eighteen calls, including those for aerial predators and food. Discussing the predator alarm call across species, Marler writes,
action. In [in progress] we articulate an alternative view according to which expressive behavior does its work more directly. A cub‘s cry of hunger enjoins: Feed me! A shriek of fear: Flee this! Relevant here is Millikan (1995), though our proposal departs from hers in important ways.
24 In sophisticated and experimentally rigorous work, Tomasello and colleagues (Tomasello 2008 and references therein) offer an account of human language development inspired by the Gricean paradigm. Bar-On and Green (in progress) engages this work in some detail.

"If a sudden predatory threat is detected nearby, the most logical response might seem to be to dash for the nearest cover, to freeze, and above all to keep quiet. This is indeed an accurate description of a few birds…but for one striking fact. Birds that keep silent, whatever the danger, are very much in a minority. It is a mark of avian sociality that almost all birds possess alarm calls as key components in their suite of antipredator responses. (2004: 138)."

"Marler notes that in some species a given call can double as an alarm or contact call. In other species, the call contains an acoustic signature identifying the caller; in others, not. Again, some of the acoustic features of calls can be predicted by their function. For instance, Marler observes that a predator alarm call should be, and in fact typically is, designed in such a way as to make detection of the caller difficult. (It tends to be a narrowband pure tone, and pitched high in a range where aerial predators such as hawks cannot hear very well.) Indeed, many species have converged on the ―seet‖ alarm, which has precisely such acoustic properties."

"Many birds, including galliforms, use food calls, and among domestic chickens the rooster calls to females to offer morsels of food. Evans and Evans (1990, 2007) also document that roosters sometimes call deceptively, holding a twig in their beak while calling to a female. What is more, for a short period (up to 24 hours) females behave as if they remember having been deceived, and respond to deceiving males less eagerly."

"Birds often vocalize during or prior to aggression in addition to exhibiting the well-known ―head forward‖ display. The vocalization tends to be in a lower register, and the lowness of register is correlated with body size and thus RHP. This suggests that aggressive vocalizations are indices in the sense we defined above."

"Marler suggests that some calls are ―functionally referential‖, while others such as alarm and aggression calls behave more like emotional displays (2004: 175). Rather than infer that the latter are comparatively impoverished, Marler makes a fascinating suggestion."

"With careful study, we find that communication by emotional displays can be very complex, especially when prevarication is involved…Furthermore, if a bird couples a call with some kind of indexing behavior, such as head-pointing or gaze direction, a certain object or point in space or particular group member can be precisely specified: the combination adds significantly to the communicative potential of emotion-based signals. (2004: 176)"

"This is close to the picture we have been developing here. We suggest that a bird‘s alarm call is best understood not, à la McDowell, simply as an instinctive act of transmitting information about the presence of a predator. For this ignores the expressive dimension of alarm calls. Using our earlier terminology, a bird‘s alarm call m-expresses a bird‘s affective state. A hypothesis about such m-expressive acts is that they serve a distinctive purpose in the lives of social, minded creatures: to show the expresser‘s state of mind to a suitably endowed observer so as to move the observer to appropriate action. Even without head-pointing or gaze, an alarm call, for instance, is directed at a particular predator, in virtue of expressing a complex state of mind (mild/intense fear of x). But the state‘s complexity need not be construed along the lines of the complexity of propositional attitudes (viz., having a particular psychological attitude – e.g. being afraid – toward a propositional representation of a worldly state of affairs, e.g. that a bird from above is about to attack). Instead it could be understood as a non-propositional, yet still intentional affective state."

"To elaborate: states such as fear of x, anger at y, excitement about z, unlike purely phenomenal states (like distress, hunger, or fatigue), are directed at specific objects or situations, without however involving a propositional characterization of x, y, or z. (Such states may, of course,
have propositional analogues/relatives. One may be afraid that x is moving too fast, angry that y is coming too close, excited by the fact that z is about to give one a treat. But, intuitively, just as some psychological states – fatigue, hunger, thirst, general malaise – have no intentional objects, there are more purely affective states that do not have propositions as their intentional objects. It would take argument to show that all affective states must au fond be propositional attitudes.)."

"What justifies attribution of such states is their ability to explain and predict animal behavior" (Note 25).

"Further, if we accept that affective states can be directed at objects or situations – and thus be intentional– albeit without being propositional, the behavior that expresses these states can plausibly be seen as inheriting their intentional dimension.26

"Much of the behavior we recognize as expressive behavior in animals expresses affective states. If we are right, such states have a certain complexity, one aspect of which is their having an intentional object, though not propositional content. To say that a state is complex is of course not to say that it has parts or components that correspond to the dimensions or aspects of complexity. Similarly, a bit of behavior– such as an alarm call – that expresses a complex affective state may also lack composite structure. Marler‘s suggestion, which we endorse, is that a bird‘s alarm call can and often does express not only the bird‘s fear but also reveals the fear‘s intentional content. Similar points apply to calls of aggression, as well as to hens‘ food calls. Understanding this suggestion may give us some purchase on what renders expressive behavior an apt candidate for being a precursor of linguistic behavior."

"An alarm call can be seen as an act that m-expresses an animal‘s affective state, which is a state with a distinctive emotional character as well as intentional object – an emotion, such as fear,

25 See Allen and Bekoff 1999 for a fuller defense.

26 Our skeptic may, of course, insist that having affective states depends on having beliefs regarding their putative intentional objects. (See, e.g., Davidson (2001, passim).). Such a view is, however, contrary to recent findings in emotion research. For further discussion see, e.g., Griffiths (1998).

24

"anger, affection, directed at a specific object or situation. We can perhaps think of such acts as originating in something akin to a gasp of fear and subsequently becoming gradually both more stylized and more refined, so as to culminate in the range of calls described by Marler, which have distinct, learnable acoustic profiles, distinguishable according to broad types of threats occasioning them.27 Thus, a hawk alarm call is a great deal like ―Hawk!‖ uttered fearfully (where ―Hawk!‖ is understood as a one-word sentence, with no structure, and with sub-propositional semantic content). But alarm calls can be dissociated from the affective states that are m-expressed in acts of producing them. (Indeed, the vocal patterns can be reproduced in the absence of the original expresser.28) They can potentially gain currency as ‗stand-ins‘ for the different intentional objects of the states characteristically m-expressed when producing them (that is, different sources of threat – e.g., leopards vs. hawks vs. …). They can become vocal ‗signatures‘: distinctive, repeatable, vocal patterns, part of a repertoire of expressive vehicles."

Note their apt references to 'sub-propositional content', if nice.