Knowledge transfer without knowledge? The case of agentive metaphors in biology

https://doi.org/10.1016/j.shpsa.2018.05.002Get rights and content

Highlights

  • Three arguments for the absence of knowledge transfer in scientific metaphors are examined.

  • Metaphor success criteria in linguistics and in science are found to differ.

  • Some agentive metaphors in biology allow for knowledge transfer.

  • Intentional agent metaphors are flexible but scientifically unsatisfactory.

  • Rational agent metaphors are heuristically fruitful and guide the exploration of their scope.

Abstract

Are scientific metaphors dispensable shortcuts that encapsulate knowledge but can always be translated back? Or do they constitute cases of knowledge transfer, even if seemingly based on scientifically underdeveloped domains? This paper defends the latter view. By drawing on the linguistic theories of metaphors, we assess a variety of agentive metaphors that pervade biology. Intentional metaphors are found unsatisfying because their use is either rigid or too widely flexible. By contrast, rational agent metaphors constitute good scientific metaphors, displaying flexible use and heuristic fruitfulness. Their range of application constantly evolves because they provide guidelines that permit the exploration of the applicability of source domain in a target domain. Their unique heuristic value makes them akin to research programs and allows for knowledge transfer, because they are based on a proper scientific source domain rather than on a folk or underdeveloped one.

Introduction

Science is replete with metaphorical talk. Metaphors typically allow one to formulate statements in one domain by using vocabulary from another domain. Scientific fields often borrow concepts, methods and techniques from external sources; most such transfers stem from an underlying transfer of formal models. However, many scientific metaphors have no such formal basis and rely only on similarities with everyday concepts. How should these metaphors be understood? Do they merely offer convenient reformulations of scientific statements, as in popular science, or do they more essentially contribute to scientific progress? Answering this is not easy, due to the lack of a formal or analytical theory of metaphors. Moreover, non-model-based metaphors seem to carry a ‘literary’ aura that renders them suspect of being plainly unscientific.

More explicitly, scientific metaphors may be thought not to involve knowledge transfer (call this the no-knowledge-transfer thesis, or NKT thesis) for at least three reasons. First, scientific metaphors may be based on a domain which contains no knowledge to be transferred. Second, metaphors may provide mere shortcuts in expression, which modify the way in which it is presented but involve no transfer. Third, scientific metaphors may be understood as similar to regular metaphors in language, the effects of which are more attentional than epistemic. Overall, metaphors threaten to be little more than rhetorical or pedagogical devices.

However, such a negative verdict would be premature. In this paper, we argue that some scientific metaphors do involve knowledge transfer. We focus on a specific example: the family of hoary, pervasive agentive metaphors in biology, which consider nature or organisms as intentional or even rational agents. We describe and analyse these metaphors and their key characteristics as well as the nature of their scientific import and issue verdicts regarding their quality. As a result, we become able to quell the three aforementioned doubts. Importantly, our analysis partly draws inspiration from the linguistic theories of metaphor. Although scientific metaphors turn out to differ from linguistic ones, the latter provide an effective foil to the former and help illuminate the conditions for their success. Overall, the literary aura of scientific metaphors turns out to pave the way for understanding some of their roles in science.

The paper unfolds as follows. The first two sections introduce two views that oppose the NKT thesis. Section 2 introduces two arguments in favour of the no-knowledge-transfer thesis, in particular the view according to which biological metaphors serve as useful but dispensable shortcuts for longer explanations. Section 3 introduces the linguistic analysis of metaphors and its apparent support to the NKT thesis. The following sections explore four agentive metaphors in science in this light: nature as an intentional agent, biological entities as intentional agents (Section 4), nature as a rational agent (Section 5) and biological individuals as rational agents (Section 6). The first one provides a mere example. The second one is an apparently useful metaphor, which displays desirable characteristics according to the linguistic perspective without being scientifically satisfying. The last two cases exemplify good metaphors and allow us to pinpoint the way in which science may benefit from them: when they suggest new directions of research and guidelines for exploring them. Good scientific metaphors are akin to research programs.

Before we begin, a number of remarks are in order. First, note that this paper chiefly addresses the philosophers concerned with the role metaphors can play in science in general. However, it may also be of interest to those specifically interested in agentive metaphors, as well as in their use and impact in biology. Second, we do not intend to characterise metaphors in general, to define them analytically or to provide an exhaustive list of their scientific roles.1 Rather, we aim to counter the NKT thesis by investigating a number of related metaphors that are widespread in evolutionary biology and that turn out to involve knowledge transfer, or so we argue.

Section snippets

Agentive metaphors and the shortcut argument

Scientifically fruitful metaphors usually involve the transfer of methods or models from one developed field to another. In the billiard model of ideal gases, for instance, considering gases as formed of particles that move and collide like billiard balls leads to a variety of laws of the kinetic theory of gases. More recently, consider the increasingly popular approach labelled ‘econophysics’, in which physics methods and models are used in economics to solve problems pertaining to complexity,

Scientific metaphors and the linguistic approach

  • a

    The cognitive linguistics approach

In this section, we present an account of metaphor which builds on Lakoff and Johnson's famous cognitive linguistics approach. We then argue that it is a legitimate framework for understanding scientific metaphors,5

Intentional metaphors

We now introduce intentional metaphors, which allow us to suggest that scientific metaphors enjoy stricter criteria for success than regular metaphors, and thus to argue that the former may allow for knowledge transfer even if the latter do not. Our secondary aim is to pave the way for the discussion of rational agent metaphors in biology, which we will showcase as good scientific metaphors.

Dennett questions nature-as-intentional-agent metaphor as follows:

the chief beauty of the Darwinian

Nature as a rational agent

The metaphor of nature as an intentional agent is inseparable from the very motivations for the theory of natural selection, as we have seen. However, this is not true of another, more sophisticated metaphor, which considers nature as a rational or maximizing agent. The two metaphors partly overlap and so may easily be conflated. If natural selection is considered akin to an intentional agent because it produces living entities that ‘appear designed’, and if design is interpreted as involving a

Biological individuals as rational agents

Biology has long witnessed a tendency to treat not only nature but also biological individuals as rational agents. But how distinct are these metaphors? If natural selection tends to maximise some population-level or collective fitness, will it not produce individuals that act so as to maximise their fitness as well? The answer is not straightforward, and there are various reasons to keep the metaphors separate here. First, from the logical point of view, the fitness mean or variance of a

Agentive metaphors as research programs

We may now assess the nature and specificity of agentive metaphors in biology. It should be clear that except for the nature-as-intentional agent one, agentive metaphors exemplify some properties of metaphors highlighted by the linguistic analysis. They operate a partial transfer of features from a source domain (naïve psychology, rational choice theory) to a target domain (biology). Moreover, as the previous sections have shown, they enjoy great flexibility: there are used in different

Conclusion

Some scientific metaphors, although deeply entrenched, appear to be mere shortcuts that encapsulate knowledge but do not transfer it from one domain to another, or to stem from domains which contained little knowledge to begin with. Drawing on the linguistic analysis of metaphors, we identified several key characteristics of metaphors – partial transfer of features, flexibility of use – which once properly refined also help identify good metaphors in science. Focusing on agentive metaphors in

References (69)

  • D. Bailer-Jones

    Models, metaphors and analogies

  • C. Bicchieri

    Should a scientist abstain from metaphor?

  • J. Birch

    Hamilton's rule and its discontents

    The British Journal for the Philosophy of Science

    (2014)
  • J. Birch et al.

    Kin selection and its critics

    BioScience

    (2015)
  • A. Bourke

    The value and validity of inclusive fitness theory

    Proceedings of the Royal Society B

    (2011)
  • A. Bourke

    Principles of social evolution

    (2011)
  • R. Boyd

    Metaphor and theory Change: What is 'metaphor' a metaphor for?

  • M. Bradie

    Science and metaphor

    Biology and Philosophy

    (1999)
  • S. Cavell

    Aesthetic problems of modern philosophy

  • R. Dawkins

    The blind watchmaker

    (1986)
  • Dawkins R. (2006) [1976]. The selfish gene. Oxford: Oxford University...
  • D. Dennett

    The intentional stance

    (1987)
  • A.W.F. Edwards

    The fundamental theorem of natural selection

    Biological Review

    (1994)
  • R.A. Fisher

    The genetical theory of natural selection

    (1930)
  • D. Gentner et al.

    Metaphor is like analogy

  • S. Glucksberg

    Understanding metaphorical Comparisons: Beyond similarity

    Psychological Review

    (1990)
  • P. Godfrey-Smith

    Three kinds of adaptationism

  • A. Grafen

    Formal Darwinism, the individual-as-maximising-agent analogy, and bet-hedging

    Proceedings of the Royal Society B

    (1999)
  • A. Grafen

    The formal Darwinism project: A mid-term report

    Journal of Evolutionary Biology

    (2007)
  • A. Grafen

    The simplest formal argument for fitness optimization

    Journal of Genetics

    (2008)
  • A. Grafen

    The formal Darwinism project in outline

    Biology and Philosophy

    (2014)
  • W.D. Hamilton

    The genetical evolution of social behaviour

    Journal of Theoretical Biology

    (1964)
  • A.I. Houston et al.

    Models of adaptive Behaviour: An approach based on state

    (1999)
  • A.I. Houston et al.

    Violations of transitivity under fitness maximisation

    Biology Letters

    (2007)
  • View full text