Can a one-ounce laboratory mouse with a brain that can fit in a teaspoon really act as a model for the complex psychiatric disorders of its far larger and more neurologically complex mammalian relative Homo sapiens? And does the heritability of laboratory animal behaviours such as a preference for avoiding the open arms of an elevated maze actually tell us anything about the influence of genetics on human medical conditions such as anxiety? Questions like these have long been raised by social scientists regarding animal-behaviour genetics, and a common response has been to critically characterize such modelling approaches as deeply reductionist. Against this interpretation, Nicole Nelson strives to show, in this enlightening and lively book, that animal-behaviour geneticists are fully aware of the difficulties of drawing analogies between human and animal behaviour and take complexity very seriously.

Nelson bases her study on fieldwork conducted in the late 2000s, largely at a western American institute she anonymizes as Coast University. She remarks that she too entered the laboratory assuming that the experimental set-ups she was observing were ‘irredeemably reductionist’ (p. 7). The scientists she interviews, however, explicitly acknowledge the limitations and provisional nature of their models. What is important for them is not that models possess strict verisimilitude but rather that they can be used to establish piecemeal associations between heritable animal behaviours and targeted human psychiatric disorders. Establishing these cross-species analogies is no simple matter – the behaviours studied are often ephemeral phenomena stabilized and standardized only through controlling a vast array of often non-obvious environmental variables, including the positioning of cages and the investigator's own smell. Narratives attributing past experimental failures to inadequate engagement with complexity inform a culture of vigilance and cautious claims-making appealing to ‘epistemological modesty’ (p. 78). Nelson's animal-behavioural geneticists see such modesty not only as intrinsic to producing good science, but also as distinguishing them from what they see as the insubstantial and socially irresponsible generalizing practices attending much of human behavioural genetics, especially in light of historical controversies over race, heredity and intelligence.

Engagement with complexity by animal-behavioural geneticists is not always obvious in published works. Nelson therefore finds it more productive to look at what she calls ‘extrafactual’ work. Drawing on constructivist metaphors, she looks at how scientists use analogies from different fields to try to establish linkages between human and mouse behaviour. She terms this non-linear and piecemeal process ‘epistemic scaffolding’. Linkages may collapse under experimental scrutiny, be reinforced by data, or be supported by additional analogies. This modelling strategy copes with complexity by being provisional, modular and modifiable. These models are used to establish knowledge not only about behavioural genetics but also about the laboratory environment and its influence on behavioural phenotypes that helps scientists cope with complexity. This latter set of facts, termed by Nelson ‘epistemic by-products’, is integral to the process of knowledge production, but usually (with some important exceptions) not seen as of value as publishable knowledge. Concentrating on publicly available claims about behavioural genetics, Nelson here intimates, can lead to researchers overlooking extrafactual work and inferring a misguidedly reductionist picture of laboratory knowledge production.

In addition to highlighting the role of extrafactual work, Nelson also stresses that the partiality of animal-based models is widely recognized among behavioural geneticists. Her case study here is the mouse-based ‘nocturnal-drinking’ model used at Coast to study alcoholism. This model overcomes mouse aversion to and rapid metabolization of alcohol to produce blood alcohol content (BAC) levels comparable to those found in binging humans. But whilst these mice may ‘drink and fall over just like college students’ (‘Dr Smith’, p. 147), the notion that a lone caged mouse can model the social psychology of college student drinking seems absurd. Nelson, however, argues that these models are more ‘reproductionist’ than reductionist, seeking to simulate an aspect of a complex disorder without assuming complete analogy between model and target. Research at Coast comes to focus on the nocturnal-drinking model not because it is viewed as a complete analogue of alcoholism but rather because other models do not produce reliably high BACs or are overly reliant on compulsion. Researchers see selecting a model as inevitably involving trade-offs and recognize that it might not be the best fit for all investigations.

Nelson sees her book as an attempt to ‘expand our analytical vocabulary and make it easier to talk about the variety of facts and processes that occur in the laboratory’ (p. 209). To this end, her work is exemplary. Her study does, however, raise some further questions for historians of the life sciences interested in reductionism. Whilst she makes it clear that animal-behavioural geneticists themselves usually take complexity seriously, public discourse on the matter is still characterized by highly reductionist tropes. The bench scientists in her study prove partly complicit in promoting these narratives, as in the process of tailoring claims to public and corporate interests they will sometimes lean on deterministic narratives. Nelson finds here a ‘paradox that is researchers’ discomfort with narratives that are in many ways of their own making’ (p. 200). Whilst translating for lay audiences usually entails simplification, this might also possibly be evidence that popular media and commercial markets offer perverse incentives to scientists to downplay complexity. Even if crude reductionism is not as prevalent as feared in the behavioural-genetics laboratory, its ubiquity in society beyond the lab remains problematic.

Model Behavior is highly recommended to historians of the life sciences with a shared interest in science and technology studies. It also provides useful context on contemporary scientific practices for those interested in the history of model organisms in the study of psychiatric disorders. Philosophers of science will also find Nelson's work on modelling very relevant, particularly in terms of how modelling occurs in everyday practices.