My ongoing projects include:
Analysis of humpback whale song (Megaptera novaeangliae). I am currently working with these data in collaboration with a colleague at SUNY Buffalo. He has completed an acoustic classification of the vocalization using a Self Organizing Map, which sorts and classifies the acoustic sounds into 25 separate units or types (see Green et al, 2007). I have been working on a HAL analysis of these units, and I believe I may be seeing evidence of semantic classes (a loose analogy here might be to say these classes are like “nouns” and “verbs”). Additionally, I believe we have evidence, based examining how these semantic classes are used in relation to the geographic areas in which they were recorded, that supports current hypotheses proposing cultural dialects in several species of marine mammals (Eriksen et al, 2005; Rendell & Whitehead, 2001). I am currently working on a technique by which I can provide additional statistic support for these results.
Speech of an African grey parrot (Psittacus erithacus); Issues of inter-observer reliability. I am collaborating with a student at the University of Georgia to create a corpus of vocalizations made my an African grey parrot in a variety of conditions – for example, in the presence of her own, in the absence of the owner, and when the owner has company visiting. While transcribing the parrot’s speech from videotape, we faced the issue of establishing inter-observer reliability. I found our project presented a unique case, and because of this, I have collaborated with Robert Rosenthal on a paper that is a revise and resubmit at Animal Behaviour discussing the statistics of inter-observer reliability. Once we have established reliability, I expect the analysis of the parrot vocalizations to be less complicated than other species, as the parrot speaks English. I believe, therefore, a larger variety of analyses will be possible.
Asperger’s model mice. I have obtained, from a group at the University of California, Riverside, a set of vocalization obtained from “normal” mice (Mus spp.) and genetically altered mice used as a model for Asperger’s Syndrome. The lab which provided these vocalizations typically studies neurological aspects of the different strains of mice; however, in the course of their experiments they noticed an audible difference between their vocalizations. I am planning a simple comparison between these two sets of vocalizations. I am hoping to establish whether the difference in communication is semantic (in which case the HAL model would show contextual differences in the usage of individual sounds by the two strains), or syntactical (in which case the differences in vocalizations would stem more from the acoustic make up of the sounds). I believe this information has potential to generalize to research on communicative abilities in human Asperger’s patients, and that it may provide important insight into the types of communication deficits present in these patients.
Killer Whale Vocalizations. Lastly, I intend to work with vocalizations from wild killer whales (Orcinus orca). I have limited behavioral data from this group, and preliminary experiments have shown promising results – HAL has grouped together vocalizations commonly seen with the same behaviors (see Deecke, 2003), in addition to grouping together “aberrant” calls with their “normal” versions. The later distinction is an interesting parallel to an already established HAL phenomenon, in which the typed words “and” and “nad” are grouped together. This occurs because they are typed in identical context, despite the fact that “nad” is a typographical error (Lund & Burgess, 1996). I believe that the “aberrant” killer whale calls are a sort of vocal “typo” (such as a mispronunciation). Additionally, there is currently a hypothesis that killer whales use dialects which vary by geographic region and over the course of time (Deecke et al, 2000), and I am therefore planning to examine the vocalizations in relation to the geographic location of the animals which produced them, in the hopes of providing additional insight into this theory.
Additionally, I have been pursuing a secondary line of research dealing with novelty in animals. My interest in this stems from an intersection of two experiences - my past work as a marine mammal trainer, where I became familiar with a behavior in which an animal was asked to “do something new”, and my husband’s research on creativity in people (e.g., Kaufman, 2009). I am currently working on applying human models of creativity to animals, so that we may better describe behaviors which involve recognition of novelty, observational learning, and innovation (Kaufman & Kaufman, 2004). Despite the fact that creativity is a niche field within psychology, the work that has already been done on defining and measuring innovation in people is enormous compared to what has been accomplished in animals. Yet there is little, if any, cross-over between the two fields. In addition, because of my background in neuroscience, I have been pairing the behavioral models to neurological ones and have begun to correlate specific types of creative behaviors with specific brain regions in non-human animals. I believe that this work will influence our ideas on environmental adaption and invasive species. To illustrate by way of a very brief example, many definitions of creativity in humans require the creative product to be both novel and appropriate (building a house out of crackers may by novel, but it’s not appropriate). In the animal kingdom, this translates directly to survival – an Australian bowerbird must build a nest or do a dance novel enough to prove his fitness and attract females, however a nest too novel or a dance too enthusiastic may scare females away and is therefore inappropriate. According to some human models, what is not appropriate to the situation is not considered creative (in fact, it’s often considered crazy), and in animal models this lack of appropriateness may be seen as maladaptive for the individual bowerbird.
| Allison's Vita | Allison B. Kaufman |
|