Skip to main content
Log in

How much can we know about the causes of evolutionary trends?

  • Published:
Biology & Philosophy Aims and scope Submit manuscript


One of the first questions that paleontologists ask when they identify a large-scale trend in the fossil record (e.g., size increase, complexity increase) is whether it is passive or driven. In this article, I explore two questions about driven trends: (1) what is the underlying cause or source of the directional bias? and (2) has the strength of the directional bias changed over time? I identify two underdetermination problems that prevent scientists from giving complete answers to these two questions.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Fig. 1
Fig. 2
Fig. 3
Fig. 4

Similar content being viewed by others


  1. Wagner (1996) draws a distinction between active and passive trends that is slightly different from McShea’s passive/driven distinction. Wagner’s notion of an active trend is somewhat broader than McShea’s notion of a driven trend (see Wagner 1996, p. 992 for clarification). For purposes of this article, however, I will restrict my attention to McShea’s work. In the scientific literature, you can find discussions of active as well as driven trends, and it isn’t always clear that scientists are attending to the differences between the definitions proposed by Wagner and McShea.

  2. For more on the history of the MBL group, see Sepkoski (2005). For the original presentations of the MBL model, see Raup et al. (1973) and Raup and Gould (1974).

  3. Stanley (1973) usually gets credit for first introducing the possibility that size increase might be a passive trend. Gould (1996) offers a lovely informal presentation of the idea of a passive trend using the analogy of a drunkard staggering home from a bar.

  4. Wang (2001, p. 851) suggests that passive and driven trends represent “extremes of a continuum” and that real trends in nature usually involve some “combination of passive and driven components.” Wang has an interesting point but states it in a confusing way. The passive driven distinction is an all-or-nothing distinction, not a matter of degree, as he suggests. Either there is a bias or there is not. If so, the trend is driven. Since biases come in different strengths, there are, in a sense, different degrees of drivenness, but this does not seem to be what Wang is talking about. On the other hand, It does seem possible that a trend at a larger scale could be driven, and that it could be composed of smaller-scale subtrends that are both passive and driven. Perhaps this is what he is getting at. For example, grade inflation at a given college could be a driven trend, even though grade inflation could be a passive trend in some departments at that institution, and a driven trend in others.

  5. There is a fascinating parallel between the statistical interpretation of the passive/driven distinction and the statistical interpretation of natural selection that has recently been defended by André Ariew, Tim Lewens, Mohan Matthen, and Denis Walsh (Matthen and Ariew 2002; Walsh et al. 2002). The distinction between the two interpretations of natural selection is stated in an especially clear way by Walsh (2007). Millstein (2000) is one philosopher who has also noted parallels between the macro- and the micro-levels.

  6. Recently, Wang (2001) has further refined the subclade test so as to look at (a) skewness within subclades; (b) skewness between subclades; and (c) skewness due to changes in variance among subclades.

  7. Stanford (2001, 2006, p. 17) defines transient underdetermination as the epistemic situation in which two theories are equally well supported by all the available empirical evidence (though they may or may not be empirically equivalent). In previous work (e.g., Turner 2007, Chap. 2), I have relied on a slightly stronger conception of local underdetermination, according to which two theories or hypotheses are underdetermined when they are underdetermined transiently (in Stanford’s sense) and background theories imply that future researchers will probably never find any evidence to discriminate between them. I think the problems discussed in this article are at least serious enough to qualify as local underdetermination problems in this sense.

  8. Although Cope did say some suggestive things about body size evolution, it is probably a mistake to attribute what has come to be known as Cope’s rule to Cope himself. See Polly (1998). I myself have not been able to find any clear statements of Cope’s rule in Cope (1974).

  9. See, e.g., McFadden (1986) on horses; Arnold et al. (1995) on plankton; Jablonski (1997) on Cretaceous mollusks; Laurin (2004) on early amniotes; Van Valkenbergh et al. (2004) on North American canids; Hone et al. (2005) on dinosaurs; and Hone and Benton (2007) on pterosaurs.

  10. The straight rule is just a rule of inductive inference. If we assume a frequentist interpretation of probability, then the rule says that if the frequency of heads in a series of observed trials is h, then we may conclude that the frequency of heads in “any further prolongation of the series” will also be approximately h.

  11. The idea of a shifting upper boundary is more commonly invoked in discussions of complexity and the major transitions in evolution. Some have thought that the upper boundary on organismal complexity might jump from one place in the morphospace to another during times of major evolutionary transition, in the sense of Maynard Smith and Szathmary (1995). For discussion, see Sterelny (2007, pp. 190–194).

  12. Kleinhans et al. (2005) arrive at pretty much the same conclusion by way of a different line of argument.


Download references


I am grateful for the comments and criticisms that I received on versions of this article that I presented at the University of Pittsburgh Center for Philosophy of Science in January, 2008, and Tufts University in September, 2008. I also thank Michael Baumgartner, Delphine Chapuis-Schmitz, Richard Dawid, Mehmet Elgin, Simon Feldman, Patrick Forber, Nina Martin, Dan McShea, Sandra Mitchell, John Norton, Andrew Pessin, Ed Slowik, Kim Sterelny, and Jim Woodward for their comments on earlier versions. My work on this project was supported by a fellowship from the University of Pittsburgh Center for Philosophy of Science.

Author information

Authors and Affiliations


Corresponding author

Correspondence to Derek D. Turner.

Rights and permissions

Reprints and permissions

About this article

Cite this article

Turner, D.D. How much can we know about the causes of evolutionary trends?. Biol Philos 24, 341–357 (2009).

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: