on the dual nature of chance in evolutionary biology and paleobiology · 2018-07-03 · 1 on the...

33
On the Dual Nature of Chance in Evolutionary Biology and Paleobiology Gunther J. Eble SFI WORKING PAPER: 1998-09-085 SFI Working Papers contain accounts of scientific work of the author(s) and do not necessarily represent the views of the Santa Fe Institute. We accept papers intended for publication in peer-reviewed journals or proceedings volumes, but not papers that have already appeared in print. Except for papers by our external faculty, papers must be based on work done at SFI, inspired by an invited visit to or collaboration at SFI, or funded by an SFI grant. ©NOTICE: This working paper is included by permission of the contributing author(s) as a means to ensure timely distribution of the scholarly and technical work on a non-commercial basis. Copyright and all rights therein are maintained by the author(s). It is understood that all persons copying this information will adhere to the terms and constraints invoked by each author's copyright. These works may be reposted only with the explicit permission of the copyright holder. www.santafe.edu SANTA FE INSTITUTE

Upload: others

Post on 10-Jul-2020

1 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: On the Dual Nature of Chance in Evolutionary Biology and Paleobiology · 2018-07-03 · 1 On the dual nature of chance in evolutionary biology and paleobiology Gunther J. Eble Committee

On the Dual Nature of Chancein Evolutionary Biology andPaleobiologyGunther J. Eble

SFI WORKING PAPER: 1998-09-085

SFI Working Papers contain accounts of scientific work of the author(s) and do not necessarily represent theviews of the Santa Fe Institute. We accept papers intended for publication in peer-reviewed journals or proceedings volumes, but not papers that have already appeared in print. Except for papers by our externalfaculty, papers must be based on work done at SFI, inspired by an invited visit to or collaboration at SFI, orfunded by an SFI grant.©NOTICE: This working paper is included by permission of the contributing author(s) as a means to ensuretimely distribution of the scholarly and technical work on a non-commercial basis. Copyright and all rightstherein are maintained by the author(s). It is understood that all persons copying this information willadhere to the terms and constraints invoked by each author's copyright. These works may be reposted onlywith the explicit permission of the copyright holder.www.santafe.edu

SANTA FE INSTITUTE

Page 2: On the Dual Nature of Chance in Evolutionary Biology and Paleobiology · 2018-07-03 · 1 On the dual nature of chance in evolutionary biology and paleobiology Gunther J. Eble Committee

1

On the dual nature of chance inevolutionary biology and paleobiology

Gunther J. Eble

Committee on Evolutionary Biology

The University of Chicago

5734 South Ellis Avenue

Chicago IL 60637

In press, Paleobiology 25 (1)

*Present address: Smithsonian Institution, Department of Paleobiology, MRC-121,Washington, D.C. 20560 and Santa Fe Institute, 1399 Hyde Park Road, Santa Fe, NM 87501.Electronic mail: [email protected], [email protected]

Page 3: On the Dual Nature of Chance in Evolutionary Biology and Paleobiology · 2018-07-03 · 1 On the dual nature of chance in evolutionary biology and paleobiology Gunther J. Eble Committee

2

Abstract .- The identification of randomness and nonrandomness is a perennial problem in

evolutionary research. Stochastic thinking in evolutionary biology and paleobiology has

solidified the use of a statistical notion of chance, but the idea of chance in evolutionary

studies goes beyond statistics. A duality arises from the use of a statistical meaning, on the

one hand, and a more strictly evolutionary meaning, on the other. The former implies a

combination of indiscriminate sampling and unpredictability due to multiple causes; the

latter codifies independence from adaptation and the directionality imposed by natural

selection. Often these meanings are kept separate in evolutionary research, used in isolation

according to the empirical situation or the goal of the investigator (recognition of pattern

versus process). I argue that evolutionary studies in general and paleobiological studies in

particular can benefit from the simultaneous application of statistical and evolutionary

notions of chance. Following some background on the notion of chance and its use, I

discuss a series of examples in which insight can be gained by explicit consideration of

both meanings. Thus, typologies of extinction become clearer when phenomena like

wanton extinction are made explicit; exaptive radiations are exposed as an alternative to

adaptive radiations; the possible nonadaptive nature of deterministic chaos becomes

sensible; the nonrandomness of community-assembly is put into question; parallel

taxonomies of sorting rooted in different notions of nonrandomness are suggested as a

means of facilitating understanding of relationships across the hierarchy; developmental

constraints and self-organization are more easily distinguished from selective constraints;

and a new term, “incidentals”, is suggested to refer to both exaptations and nonaptations.

Finally, I point to ways in which the dichotomy between chance and necessity can be

approached in evolutionary theory, by showing that the dual nature of chance in evolution

entails a distinction between functional and structural necessity, and that chance ultimately

becomes a unifying concept for a number of criticisms to neo-Darwinism.

Page 4: On the Dual Nature of Chance in Evolutionary Biology and Paleobiology · 2018-07-03 · 1 On the dual nature of chance in evolutionary biology and paleobiology Gunther J. Eble Committee

3

Introduction

Stochastic approaches in evolutionary paleobiology are now widely appreciated,

after a controversial beginning (Raup et al. 1973; Van Valen 1973; Kitts 1975; Raup 1975;

Raup et al. 1975; Stanley et al. 1981). Deterministic explanations are still part and parcel of

a search for causes, but consideration of stochastic alternatives has become a desirable

feature of many studies. The view that species in particular and taxa in general may behave

as particles in space and time, and that ensemble properties and nomothetic principles are

worth pursuing through statistical analyses (Raup et al. 1973; Van Valen 1973; Raup and

Gould 1974; Raup 1977a,b; Gould et al. 1977; Gould 1980; Raup and Schopf 1978;

Schopf 1979), has been very influential and encouraged increased rigour in

macroevolutionary studies. Whatever the successes of this approach, it has shaped the way

in which randomness and nonrandomness are perceived in the field by emphasizing the

meaning and importance of null hypotheses in empirical research.

Should rejection of a random null hypothesis exempt the paleontologist from a

concern with issues of chance and luck? Or are the notions of chance and randomness

more subtle than our allegiance to the statistical paradigm would suggest? Similar

questions can be posed to neontologists (say, ecological geneticists or molecular

evolutionists) after rejection of a hypothesis of random drift or selective neutrality. The

centrality of adaptation, the randomness of mutation, and a Darwinian heritage are all

hallmarks of evolutionary studies. They are the very reasons that make the notion of

chance in evolutionary biology and paleobiology more complicated than statistics alone

would suggest.

Currently, two meanings of chance relevant to evolutionary studies exist in relative

isolation, codified in different research traditions. In addition to the established statistical

meaning, associated with sampling error and probabilistic statements (the nuts and bolts of

Page 5: On the Dual Nature of Chance in Evolutionary Biology and Paleobiology · 2018-07-03 · 1 On the dual nature of chance in evolutionary biology and paleobiology Gunther J. Eble Committee

4

quantitative empirical research), a very peculiar, "evolutionary" meaning of chance exists in

connection with the problem of the causes of variation and their Darwinian interpretation :

"random" is an attribute of phenomena that would appear to occur independently of

adaptation or ultimate utility (Simpson 1953; Mayr 1961; Gould 1982; Ridley 1993;

Marshall 1995; Futuyma 1998). Thus, mutations are random with respect to their effects,

the same being true of exaptation (Gould and Vrba 1982) and relationships across

hierarchical levels (Vrba 1983, 1989; Vrba and Gould 1986), to cite just a few examples.

This duality in meaning has not existed without unease. Comments on the

contingent and problematic status of the chance concept are often made (e.g., Gould 1982;

Vrba and Gould 1986; Ridley 1993). Previous discussions of macroevolutionary theory

(e.g., Schopf 1979; Gould 1982; Gould and Vrba 1982; Vrba and Gould 1986; Williams

1992) have addressed the issue of chance, but have not attempted to resolve the problem of

multiple meanings. This paper explicitly addresses the topic.

Here I use a series of examples across a range of research areas to highlight the fact

that the multiplicity of evolutionary problems requires somewhat different notions of

chance. Randomness in general is ultimately a relative notion (Gigerenzer et al. 1989;

Dembski 1991). While seemingly unparsimonious, a relativistic treatment of chance

reduces the danger of simplistic reliance on deterministic explanations, and enriches the

realm of empirical evolutionary research. After identifying the roots of the current state of

affairs, I point to its resolution by arguing that an evolutionary, as opposed to statistical,

concept of chance has a proper domain of applicability, and that the relationship between

evolutionary and statistical meanings is one of complementarity, not subordination. The

examples are mostly drawn from macroevolutionary topics but also include issues in

ecology and evolutionary genetics. They are meant to indicate the value of the explicit,

simultaneous use of both meanings -- a more complete appreciation for the causes of

evolutionary phenomena.

Page 6: On the Dual Nature of Chance in Evolutionary Biology and Paleobiology · 2018-07-03 · 1 On the dual nature of chance in evolutionary biology and paleobiology Gunther J. Eble Committee

5

Background

The concept of chance has changed in meaning before attaining its present, standard

statistical manifestation (see Adler 1952; Beatty 1984). Different meanings sometimes overlap

or can be viewed as derivations of other meanings, but it is instructive to identify them

separately since each contains a different emphasis. I introduce four "non-evolutionary"

notions here. Two classical positions are of limited relevance to current statistical thinking: the

view that something is random when it works in an absolutely uncaused fashion, an

indeterministic position at times associated with quantum mechanics and nonequilibrium

thermodynamics; and the view that something is random when it does not happen by virtue of a

deliberate plan, hence its unexpected character. The latter notion was popularized by natural

theology early in the nineteenth century (Paley 1831). Of more relevance here are two other

notions, which together form what is here referred to as the “statistical meaning”: (1) An event

occurs at random because it is unpredictable, due to our ignorance of causes. This "ignorance"

interpretation (dating back to Laplace), is probably the most frequently used. It is usually

associated with the assumption of probabilistic behavior and indiscriminate sampling (viewed

as a property of independent events in nature, not of experimental design; see Sokal and Rohlf

1995:p.68 for a distinction). (2) An event may be said to be random when it is the result of the

confluence of independent causal chains, as in traffic accidents (this "coincidence"

interpretation stems from Aristotle). Notion (1) is more operational, since the suspension of

search for individual causes allows us to rely on probability statements and predictions. Notion

(2) is perhaps more relevant to the study of natural processes, in its allusion to causal

pathways. In effect, these two versions of chance imply and complement each other, and stand

both as renditions of the statistical meaning of chance.

This statistical meaning is an integral part of the analytical arsenal of every evolutionary

biologist and paleobiologist, permeating parametric and nonparametric approaches, including

Page 7: On the Dual Nature of Chance in Evolutionary Biology and Paleobiology · 2018-07-03 · 1 On the dual nature of chance in evolutionary biology and paleobiology Gunther J. Eble Committee

6

resampling. It constitutes the basis of much of everyday research related to the identification of

patterns (e.g., inhomogeneities in diversity or disparity time series, phylogenetic structure,

deviations from the molecular clock). Alongside such a notion, in effect the received view of

chance in the natural sciences, another meaning is present in evolutionary biology. It dates

back to Darwin and was initially related to the problem of the origin of variability. In modern

form, it is principally through the interpretation of the phenomenon of mutation, expressed in

the dictum "mutation is random", that this meaning of chance enters evolutionary theory (Sober

1984). By extension, its implications are far-reaching, spanning all sorts of evolutionary

studies. As Beatty (1987: p. 231) puts it, "...the notion of chance variation, as originally

conceived by Darwin, continues to play a conceptually very important role in discussions of

evolutionary change." From Darwin on, including recent usage and textbook ratification, the

gist of the evolutionary notion of chance is that events are independent of an organism's need

and of the directionality provided by natural selection in the process of adaptation . While for

mutations evolutionary randomness is clear (the alternatives are Lamarckism and

predetermination), for most other evolutionary phenomena this is not trivial, given the

directionality implicit in the expectation of selection operating as a constructive, creative force

towards adaptation over many generations (Mayr 1963; Gould 1982). "Directionality" is here

materialistic and has no implication of teleology. Patterns are evolutionarily random whenever

selection and adaptation are not directly involved.

The original development of this notion of chance by Darwin is quite informative. The

emerging fields of probability theory and statistics had little influence (Beatty 1987). Further,

there is enough variability in usage in Darwin's writings to trace his evolutionary notion of

chance to an interplay of ignorance (the problem of prediction), natural theology (the problem

of design) and coincidence (the problem of ultimate sources) interpretations of chance (Beatty

1984; Hodge 1987). This does not detract from the uniqueness of the evolutionary notion -- it

is simply not reducible to the statistical notion because the latter is silent about adaptation.

Page 8: On the Dual Nature of Chance in Evolutionary Biology and Paleobiology · 2018-07-03 · 1 On the dual nature of chance in evolutionary biology and paleobiology Gunther J. Eble Committee

7

The composite nature of the evolutionary notion of chance does allow clarification

of its roots, as well as its implications. First, Darwin always acknowledged that variation,

while not directed, is not uncaused, such that natural laws of variation might be

discoverable. This is important, for it implicitly allows for biases in the production of

variation. Second, there is a clear relationship between evolutionary chance and chance in

natural theology. The absence of a design, plan or purpose is the hallmark of chance as

conceived by both Darwin and the natural theologians. In Darwin's materialistic terms, one

would be talking about the independence of generative causes (of variation) from

adaptation. In the evolutionary paradigm, function and apparent purposefulness continued

to be the primary features demanding explanation in the organic world. Whether adaptation

should have the status of default expectation is an issue in itself (Gould 1989a; Antonovics

and van Tienderen 1991), but adaptation is a central concept in evolutionary theory both

among supporters and critics.

On the Current Use of Chance in Evolutionary Biology

In light of the distinctions discussed above, the contrast between evolutionary and

statistical meanings as currently used is summarized diagrammatically in Figure 1. Each

usage brings its own subtleties, however. Noticeable are the problems of the nature of

variation and of the meaning of null hypotheses, associated with the evolutionary notion

and the statistical notion, respectively.

Evolutionary Usage . - The eclipse of developmental biology through much of the

history of Darwinism and neo-Darwinism was accompanied by a decline of interest in the

sources of variation (Kauffman 1993; Goodwin 1994). The underemphasis on the problem

of the origin of variation in the theory of natural selection has led to some confusion over

the use of the concepts of chance and random variation. Many neo-Darwinists tend to

assume that intrinsic constraints on variability are unimportant, if not absent for all practical

purposes, and accordingly discuss the fate of variation as a result of extrinsic, selective

Page 9: On the Dual Nature of Chance in Evolutionary Biology and Paleobiology · 2018-07-03 · 1 On the dual nature of chance in evolutionary biology and paleobiology Gunther J. Eble Committee

8

forces (e.g., Charlesworth et al. 1982). Perpetuation of this line of discourse has led

sometimes to a conflation of neo-Darwinism with the assumption of absolutely unbounded

variation, by supporters and critics alike. Ho and Saunders (1984a:Figure 1. Diagrammatic representation of randomness and nonrandomness in statisticaland evolutionary senses. In (a), the distribution is statistically random because the positionof particular points is not predictable (causal chains corresponding to X and Y arepotentially independent). In (b), the distribution is statistically nonrandom because theposition of particular points is statistically predictable (causal chains corresponding to Xand Y are potentially nonindependent). If X and Y are biological variables, the distributionin (a) would be evolutionarily random if the lack of association could be shown not to begrounded on, say, conflicting selection pressures. In (b), the association observed wouldbe evolutionarily nonrandom if it could be shown to be an adaptation. If the association hascauses other than natural selection (e.g., intrinsic correlations of growth), then it isevolutionarily random despite its statistical nonrandomness.

x

xx

xx

xx

x

xx

xxx

xx

x

x

x

xx

xx

x

x

x

x

xx

(A)

X

Y

x

xxxx

xx

x xx

xx x

xxx

x

x xx x

x xxx xx

x

(B)

Y

X

p.5), for example, who criticize neo-Darwinism from a structuralist perspective, have

asserted: " The neo-Darwinian concept of random variation carries with it the major fallacy

that everything conceivable is possible." The same ambiguous interpretation of the

evolutionary notion of chance is present in the writings of scientists with similar

backgrounds (e.g., Schoffeniels 1976; Fox 1984; Matsuno 1984; Wicken 1984; Lima-de-

Faria 1988), and even in the mainstream of evolutionary theorizing (e.g., Rosen 1982;

Brundin 1986). This situation is in part a result of the sometimes exaggerated neo-

Page 10: On the Dual Nature of Chance in Evolutionary Biology and Paleobiology · 2018-07-03 · 1 On the dual nature of chance in evolutionary biology and paleobiology Gunther J. Eble Committee

9

Darwinian emphasis on variability as a given, as indicated. But it also reveals the tacit

assumption that the "standard", statistical notion of chance is our only guide to pattern in

nature. A more flexible approach is needed, as discussed below.

Criticisms of neo-Darwinism that attempt to refute the notion of chance in evolution

altogether are misdirected. The evolutionary notion of chance is not in the realm of neo-

Darwinism alone (notwithstanding its use ever since Darwin), but of evolutionary biology

as a whole (which is not exclusively neo-Darwinian). As long as biases in the production

of variation are acknowledged, this notion can continue to be used fruitfully, for its ability

to underscore the centrality of the debate over the importance of adaptation. Consider for

example Wright (1967: p.117): " The meaning of 'random' ... is that the variations are, as a

group, not correlated with the course subsequently taken by evolution (which is determined

by selection). The variations are, of course, severely limited in kind by the accumulated

results of past evolution." Or Gould (1982: p.386): "By 'random' in this context,

evolutionists mean only that variation is not inherently directed towards adaptation, not that

all mutational changes are equally likely." It is important to note that this evolutionary

meaning is acknowledged and discussed in other academic circles as well (e.g., Sober

1984; Popper 1992).

Statistical Usage . - The statistical meaning of chance usually applies to the

identification , as opposed to explanation, of pattern over evolutionary time. Beatty (1984,

1992) gives a complete account of the issue by reference to population genetics and random

drift. In the search for patterns, it is a statistical notion of chance, associated with the null

hypothesis of indiscriminate sampling (under strict independence or Markovian semi-

independence of events), that underlies accounts of drift, and for that matter, a gamut of

other evolutionary and biological problems in ecology, macroevolution, paleobiology,

systematics and much of experimental biology.

This statistical notion of chance is routinely used in most of evolutionary biology,

but this is not equivalent to the use of null models, for null models are simplified

Page 11: On the Dual Nature of Chance in Evolutionary Biology and Paleobiology · 2018-07-03 · 1 On the dual nature of chance in evolutionary biology and paleobiology Gunther J. Eble Committee

10

hypotheses that do not incorporate causes of special interest and are used to test a particular

interpretation of putative orderliness in the empirical data (see Wimsatt 1987). As such,

null models can sometimes be deterministic. For example, nonstochastic exponential and

logistic growth models have been successfully tested against diversity data (Sepkoski 1991;

Eble 1998a). Of relevance here are stochastic null hypotheses about the behaviour of a set

of data, that invariably rely on a notion of randomness akin to the ignorance interpretation

of chance. This means that objective predictions in terms of probability or probability

distributions (e.g., normal, binomial, exponential) are possible (see Raup 1977, 1991), even

if proximal causes are not known and the shapes of probability distributions have uncertain

meaning.

Science as hypothesis testing is not immune to methodological criticism (e.g., Van

Valen 1976, 1985). It also faces the fact that testing protocols are inevitably dependent on

scales of perception -- of researchers and evolutionary units (Lewontin 1966; Raup and

Schopf 1978; Schopf 1979; Wimsatt 1980; Dembski 1991). This is especially relevant in

paleontology: a nonrandom pattern on a microevolutionary scale (e.g., selection within a

species) may be inconsequential in macroevolution, when clades behave as particles (e.g.,

clade drift).

Caveats aside, a common goal in evolutionary research is to show that an apparent

pattern, which is often intuitively obvious, is indeed concrete beyond an expectation based

on some theory (see Kitts 1975; Raup et al. 1975). If a random null hypothesis is not

rejected, it is the assumption of a specific pattern that is challenged, not the possibility that

some pattern may indeed be present. Parsimony leads us to acknowledge randomness as

conceivable, nothing more.

On the other hand, the possibility remains that support of random models may be

really telling us something about the way nature is organized (or disorganized). Stochastic

models commonly applied in paleontology, for example, incorporate certain constraints and

assumptions like stasis and constancy of speciation and extinction probabilities that are

Page 12: On the Dual Nature of Chance in Evolutionary Biology and Paleobiology · 2018-07-03 · 1 On the dual nature of chance in evolutionary biology and paleobiology Gunther J. Eble Committee

11

themselves indicative of a particular kind of pattern (e.g., Raup 1977). Neutral models of

molecular evolution, if found to fit the data, naturally lead to claims that selective neutrality

is the "cause" underlying the behaviour of the system. Random models in evolutionary

developmental biology, like Kauffman's (1989, 1993) "genomic regulatory networks", are

constructed under the constraint that N genes are regulated by K other genes, entailing a

series of self-organized "generic properties" to the modelled systems. The important point

is that the use of a probability-based, statistical notion of chance can lead to perceptions of

pattern whether or not the random null hypothesis is rejected.

The Simultaneous Use of Both Meanings of Chance: Examples

As implied above, in evolutionary studies it is almost always true that the

evolutionary meaning of chance is methodologically isolated in issues surrounding process,

while pattern-oriented research goes on as usual in its use of a statistical meaning. The

explicit analysis of empirical situations using both meanings, regardless of whether one is

primarily dealing with pattern or process, is a possibility rarely appreciated. Rather than

creating confusion, it may actually enrich understanding. The following examples (see

Table 1) are designed to help put this claim in focus. Unless otherwise noted, they refer to

only one focal level in the hierarchy.

(1) Extinction Studies. - Raup (1991) categorized three different types of

extinction: (1) random extinction in a statistical sense, with no selectivity

(nonconstructive); (2) selective extinction in a Darwinian (constructive) sense, with

survival of the better adapted individuals; and (3) wanton extinction, selective extinction

independent of adaptation in normal times (nonconstructive). These alternatives are

necessary to avoid tautological formulations of the extinction problem (cf. Schopf 1979).

In terms of a non-evolutionary notion of chance, the main distinction here is between

complete randomness (1) and selectivity in general (2, 3). By contrast, an evolutionary

notion of chance clearly separates selective Darwinian extinction (2) from completely

Page 13: On the Dual Nature of Chance in Evolutionary Biology and Paleobiology · 2018-07-03 · 1 On the dual nature of chance in evolutionary biology and paleobiology Gunther J. Eble Committee

12

random extinction and wanton extinction (1, 3), i.e., constructive from nonconstructive

extinction. Wanton extinction is, in the important evolutionary sense, random (Table 1).

Pronounced physical perturbation is outside the experience of organisms and their

Table 1. Examples of simultaneous use of statistical and evolutionary meanings of chancein evolutionary biology. Each evolutionary phenomenon may be random or not under bothmeanings. Question marks indicate dependence on particular empirical situations. See textfor discussion.

Page 14: On the Dual Nature of Chance in Evolutionary Biology and Paleobiology · 2018-07-03 · 1 On the dual nature of chance in evolutionary biology and paleobiology Gunther J. Eble Committee

13

Chance concept

Statistically random

Evolutionarily random

Evolutionary phenomenon

Darwinian extinction

Wanton extinction

Onshore ordinalorigination bias

Evolutionary radiations

Deterministic chaos

Community-levelproperties

Geneticdrift

Fluctuation inselection pressure

Developmentalconstraint

Exaptation/non-aptation

Self-organization

Incidental effect

Clade selection

Sorting

no

no

no

no

no

no

yes

yes

no

?

no

no

no

?

no

yes

?

?

?

?

yes

?

yes

yes

yes

yes

no

?

adaptations to local environments during background times, and in that sense is

unpredictable (Jablonski 1989). Whatever preferential survival occurs through wanton

extinction, it occurs because of fortunate cooptations of organismic features that happened

to be advantageous in the extremely perturbed environments thought to be associated with

Page 15: On the Dual Nature of Chance in Evolutionary Biology and Paleobiology · 2018-07-03 · 1 On the dual nature of chance in evolutionary biology and paleobiology Gunther J. Eble Committee

14

extinction events in the fossil record. No direct adaptation is involved. While differential

higher-level adaptation is always a plausible explanation for extinction selectivity, the dual

meaning of chance exposes wanton extinction as an equally plausible alternative. This

duality might facilitate refinement of other typologies of extinction as well, such as when

the sources of selectivity or hierarchical lines of causation are specified.

(2) Origins of Higher Taxa. - Jablonski and Bottjer (1991) demonstrated the

existence of a statistically nonrandom pattern of higher ordinal origination in onshore

environments, as opposed to offshore ones. The use of a statistical notion of chance here

shows a very intriguing pattern, and opens the way for a challenging question under an

evolutionary notion: is the pattern evolutionarily random (e.g., if driven by environmental

stresses affecting development) or nonrandom (e.g., if there is preferential survival of

novelties onshore)?

Similarly, major evolutionary radiations, such as in the Cambrian and Ordovician,

are definitely nonrandom when compared with background times, but much of the debate in

paleobiology and systematics surrounds the evolutionary randomness (intrinsic constraints)

or nonrandomness (extrinsic, selective constraints) of patterns of radiation. In other words,

the debate is about whether particular radiations are exaptive, based on developmental

flexibility or on the incidental cooptation of a conserved trait that happens to increase the

chances of speciation, or whether they are adaptive, based on true key innovations that

enhance the occupation of empty ecospace (see Eble 1998b). Notice that independence

from adaptation (evolutionary randomness) does not need to entail independence from

fitness: both exaptive and adaptive radiations imply higher emergent fitnesses. It is worth

noting that a number of "adaptive" radiations might well turn out to be exaptive, if this latter

category were more fully incorporated in our interpretive schemes.

(3) Determinism in Ecology and Paleoecology. - Population growth dynamics,

competition, and predator-prey interactions have been increasingly modelled as a chaotic

process, where a simple underlying structure (usually associated with the nature of the

Page 16: On the Dual Nature of Chance in Evolutionary Biology and Paleobiology · 2018-07-03 · 1 On the dual nature of chance in evolutionary biology and paleobiology Gunther J. Eble Committee

15

equations used to describe the process), leads to the identification of strange attractors that

justify the expression "deterministic chaos" (Schaffer 1985; May 1987). Less attention has

been paid to the evolutionary status of the attractors themselves. To the extent that the

attractors result from the inherent properties of nonlinear dynamics and its equations, and

not because of adaptation, one can say that ecological attractors are random in an

evolutionary sense. However, it has been argued that in some instances chaotic regimes

may be favoured by selection (Ferriere and Fox 1995). In this case, a claim for

evolutionary nonrandomness would be justified. More research is needed on the

relationship between chaotic patterns and evolutionary processes, but it is apparent that the

use of both statistical and evolutionary meanings of chance helps puts into focus the larger

evolutionary implications of chaos.

In turn, community-assembly and community-disassembly (Drake 1991; Pimm

1991; Mikkelson 1993) have been shown to follow certain rules, which are nonrandom

under a statistical notion of chance. The exact nature of these rules is still a matter of

debate, but in several cases they hinge on ecological processes such as competition and

predation. This suggests that they are nonrandom in an evolutionary sense as well, given

the importance that Darwinian mechanisms of adaptation and natural selection are likely to

have. If, however, what is seen as a community is in fact an assemblage of species

behaving individualistically (Underwood 1986; Jablonski and Valentine 1993; Jablonski

and Sepkoski 1996), then the whole problem collapses, and statistical and evolutionary

characterizations in terms of community behavior are not called for. The claim for

generality of paleoecological stasis (e.g., Brett and Baird 1995; Morris et al. 1995), for

example, faces such a challenge, since explanations that rely on nonrandom evolutionary

phenomena, like ecological couplings based on selection, are only meaningful if "static"

assemblages over geological time are more than filtered individualism written large. This is

still a matter of contention, for the extent to which patterns of compositional stability in the

fossil record depart from null expectations has only now begun to be more rigorously

Page 17: On the Dual Nature of Chance in Evolutionary Biology and Paleobiology · 2018-07-03 · 1 On the dual nature of chance in evolutionary biology and paleobiology Gunther J. Eble Committee

16

explored (Baumiller 1996; Bennington and Bambach 1996; Holland 1996). Claims for

determinism in ecology and paleoecology, and the application of chance notions, must be

tied to the identification of ecological and evolutionary units at the appropriate hierarchical

level.

(4) Hierarchy Theory. - In the context of hierarchy theory, relationships across

levels can become more explicit when both notions of chance are used. To follow up on

the example above, if community-level properties were the blind by-product of processes at

lower levels, one would have evolutionary randomness even in the face of nonrandomness

at lower levels. However, if it could be shown that community-level selection is taking

place (Wilson 1988), those properties would not be random with respect to adaptation at

that level.

In macroevolution, clade sorting, being just differential origination or extinction,

can be random or not in both senses (Table 1). Vrba and Gould (1986) classified

nonrandom causes of sorting in a statistical sense only, arguing that a hierarchical approach

based solely on a statistical notion of chance avoids the relegation of deterministic causes of

phenomena above or below the focal level of putative selection to the black box of

randomness (e.g., species selection above and segregation distortion or substitution bias

below the focal level of populations) . While this stance highlights deterministic controls

underlying multilevel interactions, it may obscure the often incidental nature of such

interactions. Incidental effects (Vrba 1983, 1989) are random in an evolutionary sense, and

this is a dimension that deserves equal consideration in hierarchical studies. From a

mechanistic perspective, underscoring independence from adaptation (Darwinian

indeterminism) or the existence of deterministic controls beyond selection and adaptation

has the same heuristic value -- they are complementary facets of an empirical problem. The

investigation of particular causes behind multilevel processes need not commit to a

statistical notion of nonrandom sorting, if only because it is the relative importance of

hierarchical selection (and thus adaptation) what we want to determine. The simultaneous

Page 18: On the Dual Nature of Chance in Evolutionary Biology and Paleobiology · 2018-07-03 · 1 On the dual nature of chance in evolutionary biology and paleobiology Gunther J. Eble Committee

17

use of both notions of chance can be instrumental in achieving this objective, by making

more explicit the ways in which selection may or may not be involved in hierarchical

phenomena. This by itself reduces the risk of underemphasis on particular aspects of

biological causality, and enriches the realm of hierarchical explanations.

Thus, when clade sorting is categorized either in terms of incidental effects (Vrba

1983, 1989) or clade selection, a distinction arises, for while neither is statistically random

in the absence of clade drift, the former is evolutionarily random but the latter is not (Table

1). By the same token, in the generative aspect of clade sorting, origination may be

statistically random (no birth bias) or not, but in both instances it will be evolutionarily

random as a reflection of Wright's rule (Wright 1967): origination is blind with respect to

potential long-term directionality supposedly provided by natural selection. In Williams'

(1992) terms, gene pool divergence occurs without regard to its significance for the

persistence of the changed gene pool. Even so, origination may be evolutionarily

nonrandom below the focal level, if its immediate cause happens to be natural selection.

Thus, the scale of the problem at hand may determine what is random or not, and in what

sense. Generally speaking, upward or downward effects that lead to increased fitness at the

focal level will tend to be evolutionarily random, and only emergent features at a given focal

level will be subject to a claim for evolutionary nonrandomness, regardless of how one

defines processes (e.g. species selection) at the focal level. Species selection may be

defined in terms of emergent fitnesses instead of emergent characters (Lloyd and Gould

1993; Grantham 1995), but whenever emergent fitnesses are incidental consequences of

processes at different levels, "selection" will be random with respect to adaptive value at the

focal level. The complexity of multilevel relationships (see Vrba 1989) should benefit from

the mutually informative character of parallel taxonomies of sorting rooted in different

perceptions of chance and, by extension, nonrandomness.

(5) Evolutionary Genetics. - How random is genetic drift? The notion of drift

simultaneously incorporates both meanings of chance. Indiscriminate sampling can arise

Page 19: On the Dual Nature of Chance in Evolutionary Biology and Paleobiology · 2018-07-03 · 1 On the dual nature of chance in evolutionary biology and paleobiology Gunther J. Eble Committee

18

by either sampling error or selective neutrality (which occurs even in the absence of small

sample sizes). In either case, there is also no role played by adaptation. The evolutionary

definition of chance is tacitly accepted here, which helps explain why genetic drift has

become a paradigm of randomness in evolution. At the limit, "drift would be stochastic and

selection deterministic even if quantum mechanics were someday discarded and replaced by

a deterministic theory" (Sober 1984: p.115).

Another situation in which the use of an evolutionary notion of chance helps in

understanding process is the nature of random fluctuations in selection pressure. By

including fluctuations in systematic pressures under the category of (stochastically) random

fluctuations, Wright (1949 and later; see Beatty 1992) has blurred the distinction between

genetic drift and natural selection. Regardless of how one views this dilemma, random

variation in selective pressures is nonrandom from an evolutionary perspective, since it is

coupled with variation in expected optimal adaptations at the focal level. However,

fluctuations in selective pressures may often be random with respect to higher-level

adaptation, as a result of Wright's rule.

(6) Evolutionary Developmental Biology. - Developmental constraints (biases in

the production of variation caused by inherent properties of the developmental system -- see

Maynard Smith et al. 1985) exist in relative independence from the environment. It is thus

clear how the notion is distinct from selective constraint (stabilizing selection), regardless of

temporal scale and past history: developmental constraints are random in the sense of

prospective independence from adaptation. Putative developmental constraints must be

recognized as instances of either exaptation or nonaptation, following Gould and Vrba's

(1982) terminology. If the resulting structures are functional, they are exaptations; if

nonfunctional, they are nonaptations. In any case, "...all exaptations originate randomly

with respect to their effects" (Gould and Vrba 1982: p.12); the same is true for

nonaptations. To the extent that exaptations and nonaptations are conceptually alike, one is

faced with the possibility of finding a term that can refer to both (see Gould and Vrba

Page 20: On the Dual Nature of Chance in Evolutionary Biology and Paleobiology · 2018-07-03 · 1 On the dual nature of chance in evolutionary biology and paleobiology Gunther J. Eble Committee

19

[1982: p.12] on the need for terms unburdened by "convictions about the supremacy of

adaptation"). I suggest that, granted the insight of an evolutionary meaning of chance, both

exaptations and nonaptations can be called biological incidentals . The term "incidentals" is

preferable over "fortuities" or "accidents" because it avoids some of their positive and

negative connotations, respectively. An incidental is a feature that appeared for reasons

unrelated to current functionality, and therefore independently of current adaptation,

whether or not a function exists. "Spandrels" (Gould and Lewontin 1979; Gould 1997)

represents a related term, but it refers only to architectural by-products, which could be

viewed as a subset of incidentals. Thus, all developmental constraints are in principle

biological incidentals. This realization may throw light on the supposed "chaos of

constraint terminology" (Antonovics and van Tienderen 1991), by unambiguously

conveying a distinction from selective constraints and reducing the overlap among different

classes of constraint.

Self-organization in ontogeny is another concept that has to be viewed in different

light under an evolutionary meaning of chance. Kauffman (1993, 1995) and Goodwin

(1994) clearly distinguish between self-organization and selection. While both are involved

in generating order, the distinction holds because self-organization implies spontaneous

order, arising from inherent properties conducive to ahistorical universals. The distinction

is made even clearer when one recognizes that self-organized order is logically independent

of adaptation (although it may and probably usually does interact with it), and therefore

random in evolutionary terms. Oxymoronic as it may sound, to speak here of "random

order" (in this case, order from intrinsic properties) is clarifying for it marks a distinction

from "selective order" (externally determined order). The origins of evolutionary order and

the hopes for a marriage of self-organization and selection are thus put in proper

perspective.

Page 21: On the Dual Nature of Chance in Evolutionary Biology and Paleobiology · 2018-07-03 · 1 On the dual nature of chance in evolutionary biology and paleobiology Gunther J. Eble Committee

20

Although some of these distinctions are self-evident, many are rarely explicitly

recognized, with the potential for misunderstanding of the basic issues that surround

several ongoing controversies in evolutionary biology. Randomness and chance are

relative notions, i.e., they are dependent on the particular meaning used. Different

meanings are context-specific, and failure of specification can easily lead to context-

dependent ambiguities. This is also true of fitness (Sober 1984). Beyond the clarification

of individual topics, the simultaneous use of both meanings of chance in a systematic way

could be instrumental in addressing the more general problem of the role of chance in

evolution, with consequences for the structure of evolutionary theory.

Chance and Necessity

More than twenty-five years ago, Jacques Monod (1972) surprised the scientific

community with his treatment of the old Democritan distinction between chance and

necessity: chance was turned into effectively unbounded random variation, necessity into

selection (Fig. 2, topology #1). "Drawn out of the realm of pure chance, the accident enters

into that of necessity, of the most implacable certainties. For [then] natural selection

operates ..." (Monod 1972: p. 118; emphasis added). Although the distinction was clearly

within the Darwinian framework, it was considered simplistic (e.g., Dobzhansky 1974),

and it failed to underscore the difference between evolutionary and non-evolutionary

notions of chance, as discussed in this paper (although the distinction was mentioned for no

effect in passing). Also, it failed to account for the difference between two kinds of

necessity: functional necessity (adaptation) and structural necessity (natural or physical

necessity, but with no implication of absolute inevitability; cf. Popper 1959, Beatty 1995).Figure 2. A simplified logical scheme for the status of chance in evolutionary theory.Different positions are represented in a transformational series, going from Monod'sdistinction between chance (mutation) and necessity (selection) (topology #1), to therecognition of two different kinds of necessity, with structural necessity clustered withstatistical chance (topology #2), to a simple dichotomy of functional necessity, representingthe neo-Darwinian core of adaptation and natural selection, as opposed to chance underboth meanings, whereby structural (natural) necessity is rendered random in anevolutionary sense and chance thus stands as a unifying concept for a variety of

Page 22: On the Dual Nature of Chance in Evolutionary Biology and Paleobiology · 2018-07-03 · 1 On the dual nature of chance in evolutionary biology and paleobiology Gunther J. Eble Committee

21

developments lying at the periphery of neo-Darwinism (topology #3). See text for furtherdiscussion.

TOPOLOGY #1EVOLUTION

N ECES SIT Y C H ANC E

TO POLOGY #2

FUN CT ION AL N ECES SIT Y

ST RU CT UR AL NEC ESSIT Y

ST AT IST ICALCH ANC E

EVOLUTION

INCIDENTALS (EXAPTATION, NONAPTATION)ADAPTATION

DETERMINISM INDETERMINISM

EVOLUTION

FUN CT ION AL N ECES SIT Y CH ANC E

NEO-DARWINIAN CORE NEO-DARWINIAN PERIPHERY

TOPOLOGY # 3

Functional necessity is adaptation-oriented. Structural necessity is blind,

spontaneous, automatic. Since Darwin, the essence of the evolutionary perception of

chance versus necessity has been the opposition of adaptation and the rest (Hodge 1987).

It is the evolutionary notion of chance that makes such opposition sensible, by allowing

Page 23: On the Dual Nature of Chance in Evolutionary Biology and Paleobiology · 2018-07-03 · 1 On the dual nature of chance in evolutionary biology and paleobiology Gunther J. Eble Committee

22

structural necessity to be lumped with statistical chance (Fig. 2, topology #2). This clearly

demarcates the realm of biological incidentals (exaptations, nonaptations) in terms of

nondarwinian evolutionary mechanisms (deterministic or not). At the limit, the distinction

that evolutionists really face is between chance (under both meanings) and functional

necessity (Fig. 2, topology #3). This constitutes a simple way of summarizing the variety

of debates that relate to the adequacy of neo-Darwinism. The emphasis on functional

necessity is the core of neo-Darwinism. Most challenges to this core, which can be said to

lie in the periphery of neo-Darwinism, are inextricably linked to the perception that chance

(under either meaning) is bound to be important in evolution. There is thus a connection,

through the dual nature of chance, among the various criticisms to neo-Darwinism, a

suggestion that deserves further scrutiny.

Needless to say, the dichotomies here discussed are not intended to encapsulate all

aspects of a supposedly binary reality. They are just instruments for further thought. In

particular, the account of chance in this paper is distinct from the so called "evolutionary

contingency thesis" (Beatty 1995). Contingency is a subtle concept, at the border of chance

and necessity (Gould 1989b), and with obvious importance to the understanding of

evolution. Contingency is the confluence of all causal chains; individual causes may at

times reflect functional necessity, but because of Markovian constraint, accidental events

and boundary conditions (Bambach 1996), the historical network of causation may be more

an expression of chance (under both meanings) than functional necessity. Still, due to its

hybrid character, it is unclear where exactly contingency lies in between the nomothetic

(search for laws) and idiographic (uncovering history) halves of evolutionary (paleo)

biology. Its status in evolutionary theory should be further refined.

Conclusions

The use of a statistical notion of chance helps us identify pattern. The use of the

evolutionary notion, in turn, helps us interpret pattern in evolutionary fashion. Evolutionary

phenomena can often be rendered random by application of the latter. It has been seen that

Page 24: On the Dual Nature of Chance in Evolutionary Biology and Paleobiology · 2018-07-03 · 1 On the dual nature of chance in evolutionary biology and paleobiology Gunther J. Eble Committee

23

this is not unfair. In a paradoxical but meaningful way, "random order" is fully possible in

evolutionary biology. In other words, one has "order for free" (Kauffman 1993, 1995),

without purpose (Goodwin 1994).

Ironically, it is still possible to see such an evolutionary indeterminism being similar

to the more common notion of indeterminism in the nonbiological sciences, that is, chance

as unpredictability. In evolutionary terms, one can say that a phenomenon is unexpected

when it does not occur in conformity with the prediction of orientation (via natural

selection) toward adaptation. In addition, evolutionarily random or nonrandom events can

certainly be redescribed in terms of the confluence or not of independent causal chains. The

intent here, however, is not to deny an in principle reducibility of chance concepts to a

single notion, whatever it may be. It is instead to highlight the empirical value of duality in

meaning, at least for evolutionary purposes. Evolutionary biology and paleobiology deal

with mechanisms that are absent in other natural sciences, and our conceptual schemes

should reflect this fact as much as possible. If selection did not exist, there would be no

need for an evolutionary meaning of chance. Electrons have no fitness. Biological entities

potentially have some fitness, and an important goal of evolutionary research is to

determine the relative importance of selection and adaptation in producing such entities.

Conceptual refinement is a step towards this goal.

Ultimately, one could argue, as Dembski (1991) does, that there is no absolute

notion of randomness: something is random by virtue of deviation from a theoretical

collection of patterns specified beforehand. It is the pattern of interest that will determine

what is random and what is not. This relativistic attitude is more flexible and heuristically

more appropriate in light of the complexity of evolutionary problems.

In evolution, chance means simply that: no adaptation, no selection at the focal level of

causation of evolutionary phenomena. But evolutionary studies also engage in a non-

evolutionary, statistical usage. One should view such usages as complementary, and not in

terms of subordination or as pertaining to separate empirical domains. Together they

Page 25: On the Dual Nature of Chance in Evolutionary Biology and Paleobiology · 2018-07-03 · 1 On the dual nature of chance in evolutionary biology and paleobiology Gunther J. Eble Committee

24

sharpen our thinking, and end up giving more insight and reducing confusion in a field as

conceptually fluid as evolutionary biology. Much as with "survival of the fittest", "survival

of the luckiest" (Kimura 1989) may be reasonable in more than one sense.

Acknowledgements

I thank J. Alroy, R. K. Bambach, D. H. Erwin, M. Foote, J. Huss, D. Jablonski, R.

Lupia, D. McShea, G. Mikkelson, K. Saalfeld, J. J. Sepkoski, Jr., S. Suter, L. Van Valen,

G. P. Wagner and W. Wimsatt for discussion and/or comments on different versions of the

manuscript. I also thank R. B. Sansom and an anonymous reviewer for additional

comments and suggestions. This research benefited from CNPq grant 201542/91-9, a

William Rainey Harper Fellowship (University of Chicago), the Division of Biological

Sciences of the University of Chicago, and postdoctoral fellowships from the Smithsonian

Institution and the Santa Fe Institute.

Literature Cited

Adler, M. J., ed. 1952. The great ideas I. A syntopicon. University of Chicago Press,

Chicago.

Antonovics, J. and P. H. van Tienderen 1991. Ontoecogenophyloconstraints? The

chaos of constraint terminology. Trends in Ecology and Evolution 6: 166-168.

Bambach, R. K. Progress by chance. Geological Society of America Abstracts with

Program 28: A-178.

Baumiller, T. K. 1996. Exploring the pattern of coordinated stasis: simulations and

extinction scenarios. Palaeogeography, Palaeoclimatology, Palaeoecology 127: 135-

145.

Beatty, J. 1984. Chance and natural selection. Philosophy of Science 51: 183-211.

Page 26: On the Dual Nature of Chance in Evolutionary Biology and Paleobiology · 2018-07-03 · 1 On the dual nature of chance in evolutionary biology and paleobiology Gunther J. Eble Committee

25

_______. 1987. The probabilistic revolution in evolutionary biology -- an overview.

Pp. 229-232 in L. Krüger, G. Gigerenzer, and M. S. Morgan, eds. The

probabilistic revolution. Vol. 2: Ideas in the sciences. MIT Press, Cambridge, Mass.

_______. 1992. Random drift. Pp. 273-281 in E. F. Keller and E. A. Lloyd, eds.

Keywords in evolutionary biology. Harvard University Press, Cambridge, Mass.

_______. 1995. The evolutionary contingency thesis. Pp. 45-81 in G. Wolters and

J. G. Lennox, eds. (in collab. with P. McLaughlin) Concepts, theories, and

rationality in the biological sciences. Universitätsverlag Konstanz and University of

Pittsburgh Press, Konstanz, Pittsburgh.

Bennington, J. B. and Bambach, R. K. 1996. Statistical testing for paleocommunity

recurrence: are similar fossil assemblages ever the same? Palaeogeography,

Palaeoclimatology, Palaeoecology 127: 107-133.

Brett, C. E. and Baird, G. C. 1995. Coordinated stasis and evolutionary ecology of

Silurian to Middle Devonian faunas in the Appalachian Basin. Pp. 285-315 in D. H.

Erwin and R. L. Anstey, eds. New approches to speciation in the fossil record.

Columbia University Press, New York.

Brundin, L. Z. 1986. Evolution by orderly stepwise subordination and largely

nonrandom mutations. Systematic Zoology 35: 602-607.

Burian, R. M. 1988. Challenges to the evolutionary synthesis. In M.K.Hecht and B.

Wallace, eds. Evolutionary Biology 23: 247-269. Plenum, New York.

Charlesworth, B., R. Lande, and M. Slatkin. 1982. A neo-Darwinian commentary on

macroevolution. Evolution 36: 474-498.

Dembski, W. A. 1991. Randomness by design. Noûs 25: 75-108.

Dobzhansky, T. 1974. Two contrasting world views. Pp. 131-141 in J. Lewis, ed.

Beyond chance and necessity. Humanities Press, Atlantic Highlands, N. J.

Drake, J. A. 1990. Communities as assembled structures: do rules govern pattern?

Trends in Ecology and Evolution 5: 159-164.

Page 27: On the Dual Nature of Chance in Evolutionary Biology and Paleobiology · 2018-07-03 · 1 On the dual nature of chance in evolutionary biology and paleobiology Gunther J. Eble Committee

26

Eble, G. J. 1998a. Diversification of disasteroids, holasteroids and spatangoids in the

Mesozoic. Pp. 629-638 in R. Mooi and M. Telford, eds. Echinoderms San

Francisco. A. A. Balkema Press, Rotterdam, Brookfield.

_______. 1998b. The role of development in evolutionary radiations. In M. McKinney,

ed. Biodiversity dynamics: turnover of populations, taxa, and communities.

Columbia University Press, New York.

Ferriere, R. and Fox, G. A. 1995. Chaos and evolution. Trends in Ecology and

Evolution 10: 480-485.

Fox, S. W. 1984. Proteinoid experiments and evolutionary theory. Pp. 15-60 in

Ho and Saunders 1984b.

Futuyma, D. J. 1998. Evolutionary Biology. Third edition. Sinauer, Sunderland.

Gigerenzer, G., Z. Swijtink, T. Porter, L. Daston, J. Beatty, and L. Krüger. 1989. The

empire of chance. Cambridge University Press, Cambridge.

Goodwin, B. 1994. How the leopard changed its spots: the evolution of complexity.

Scribner, New York.

Gould, S. J. 1980. The promise of paleobiology as a nomothetic, evolutionary

discipline. Paleobiology 6: 96-118.

_______. 1982. Darwinism and the expansion of evolutionary theory. Science 16:

380-387.

_______. 1989a. A developmental constraint in Cerion , with comments on the

definition and interpretation of constraint in evolution. Evolution 43: 516-539.

_______. 1989b. Wonderful life. Norton, New York.

_______. 1997. The exaptive excellence of spandrels as a term and prototype.

Proceedings of the National Academy of Sciences 94: 10750-10755.

Gould, S. J. and Lewontin, R. C. The spandrels of San Marco and the Panglossian

paradigm: a critique of the adaptationist programme. Proceedings of the Royal Society

of London B 205: 581-598.

Page 28: On the Dual Nature of Chance in Evolutionary Biology and Paleobiology · 2018-07-03 · 1 On the dual nature of chance in evolutionary biology and paleobiology Gunther J. Eble Committee

27

Gould, S. J. and E. S. Vrba. 1982. Exaptation -- a missing term in the science of form.

Paleobiology 8: 4-15.

Grantham, T. A. 1995. Hierarchical approaches to macroevolution: recent work on

species selection and the "effect hypothesis". Annual Review of Ecology and

Systematics 26: 301-321.

Ho, M.-W. and P. T. Saunders, eds. 1984a. Pluralism and convergence in evolutionary

theory. Pp. 3-12 in Ho and Saunders 1984b.

_______. 1984b. Beyond neo-Darwinism. Academic Press, London.

Hodge, M. J. S. 1987. Natural selection as a causal, empirical, and probabilistic theory.

Pp. 233-270 in L. Krüger, G. Gigerenzer, and M. S. Morgan, eds. The

probabilistic revolution. Vol. 2: Ideas in the sciences. MIT Press, Cambridge, Mass.

Holland, S. M. 1996. Recognizing artifactually generated coordinated stasis: implications

of numerical models and strategies for field tests. Palaeogeography,

Palaeoclimatology, Palaeoecology 127: 147-156.

Jablonski, D. 1989. The biology of mass extinction: a palaeontological view.

Philosophical Transactions of the Royal Society of London B 325: 357-368.

Jablonski, D. and D. Bottjer. 1991. Environmental patterns in the origins of higher taxa:

the post-Paleozoic fossil record. Science 252: 1831-1833.

Jablonski, D. and Sepkoski, J. J., Jr. 1996. Paleobiology, community ecology, and

scales of ecological pattern. Ecology 77: 1367-1378.

Jablonski, D. and J. W. Valentine. 1993. Fossil communities: compositional variation at

many time scales. Pp. 341-349 in R. E. Ricklefs and D. Schluter, eds. Species

diversity in ecological communities: historical and geographical perspectives.

University of Chicago Press, Chicago.

Kauffman, S. A. 1989. Origins of order in evolution: self-organization and selection.

Pp. 67-88 in B. Goodwin and P. Saunders, eds. Theoretical biology: epigenetic and

evolutionary order from complex systems. Edinburgh University Press, Edinburgh.

Page 29: On the Dual Nature of Chance in Evolutionary Biology and Paleobiology · 2018-07-03 · 1 On the dual nature of chance in evolutionary biology and paleobiology Gunther J. Eble Committee

28

_______. 1993. The origins of order. Oxford University Press, New York.

_______. 1995. At home in the universe. Oxford University Press, New York.

Kimura, M. 1989. The neutral theory of molecular evolution and the world view of the

neutralists. Genome 31: 24-31.

Kitts, D. B. 1975. Stochastic models of phylogeny and the evolution of diversity: a

discussion. Journal of Geology 83: 125-126.

Lewontin, R. C. 1966. Is nature probable or capricious? Bioscience 16: 25-27.

Lima-de-Faria, A. 1988. Evolution without selection. Elsevier, Amsterdam.

Lloyd, E. A. and Gould, S. J. 1993. Species selection on variability. Proceedings of the

National Academy of Sciences 90: 595-599.

Marshall, C. R. 1995. Darwinism in an age of molecular revolution. Pp. 1-30 in C. R.

Marshall and J. W. Schopf, eds. Evolution and the molecular revolution. Jones and

Bartlett, Sudbury, Massachusetts.

Matsuno, K. 1984. Open systems and the origin of protoreproductive units. Pp. 61-88

in Ho and Saunders 1984b.

May, R. M. 1987. Chaos and the dynamics of biological populations. Proceedings of

the Royal Society of London A 413: 27-44.

Maynard Smith, J., R. Burian, S. Kauffman, P. Alberch, J. Campbell, B. Goodwin,

R. Lande, D. Raup, and L. Wolpert. 1985. Developmental constraints and

evolution. Quarterly Review of Biology 60: 265-287.

Mayr, E. 1961. Cause and effect in biology. Science 134: 1501-1506.

Mayr, E. 1963. Animal species and evolution. Harvard University Press, Cambridge,

Mass.

Mikkelson, G. M. 1993. How do food webs fall apart? A study of changes in trophic

structure during relaxation on habitat fragments. Oikos 67: 539-547.

Monod, J. 1972. Chance and necessity. Vintage Books, New York.

Page 30: On the Dual Nature of Chance in Evolutionary Biology and Paleobiology · 2018-07-03 · 1 On the dual nature of chance in evolutionary biology and paleobiology Gunther J. Eble Committee

29

Morris, P. J., L. Ivany, K. Schopf, and C. Brett. The challenge of paleoecological stasis:

reassessing sources of evolutionary stability. Proceedings of the National Academy

of Sciences USA 92: 11269-11273.

Paley, W. 1831. Natural theology. Jonathan Leavitt, New York.

Popper, K. R. 1959. The logic of scientific discovery. Routledge, New York. (1992

reprint).

_______. 1992. Unended quest. Routledge, London.

Raup, D. M. 1975. Taxonomic survivorship curves and Van Valen's law. Paleobiology

1: 82-96.

Raup, D. M. 1977. Stochastic models in evolutionary paleontology. Pp. 59-78 in A.

Hallam, ed. Patterns of evolution. Elsevier, Amsterdam.

_______. 1991. Extinction: bad genes or bad luck? Norton, New York.

Raup, D. M. and T. J. M. Schopf. 1978. Stochastic models in paleontology: a primer.

Notes for workshop "Species as particles in space and time", held at the U.S.

National Museum, Smithsonian Institution.

Raup, D. M., S. J. Gould, T. J. M. Schopf, and D. S. Simberloff. 1975. Stochastic

models of phylogeny and the evolution of diversity: a reply. Journal of Geology 83:

126-127.

Ridley, M. 1993. Evolution. Blackwell Scientific, Boston.

Rosen, D. E. 1982. Do current theories of evolution satisfy the basic requirements of

explanation? Systematic Zoology 31: 76-85.

Schaffer, W. M. 1985. Order and chaos in ecological systems. Ecology 66: 93-106.

Schoffeniels, E. 1976. Anti-chance. Pergamon, Exeter.

Schopf, T. J. M. 1979. Evolving paleontological views on deterministic and stochastic

approaches. Paleobiology 5: 337-352.

Sepkoski, J. J., Jr. 1991. Population biology models in macroevolution. Pp. 136-156 in

N. L. Gilinsky and P. W. Signor, eds., Analytical Paleobiology. Knoxville:

Page 31: On the Dual Nature of Chance in Evolutionary Biology and Paleobiology · 2018-07-03 · 1 On the dual nature of chance in evolutionary biology and paleobiology Gunther J. Eble Committee

30

The Paleontological Society.

Simpson, G. G. 1953. The major features of evolution. Columbia University Press,

New York.

Sober, E. 1984. The nature of selection. MIT Press, Cambridge, Mass.

Sokal, R. R. and F. J. Rohlf. 1995. Biometry. W. H. Freeman, New York.

Stanley, S. M., Signor, P. W., Lidgard, S. and Karr, A. F. 1981. Natural clades differ

from "random clades": simulations and analyses. Paleobiology 7: 115-127.

Underwood, A. J. 1986. What is a community? Pp. 351-367 in D. M. Raup and D.

Jablonski, eds. Patterns and processes in the history of life. Springer, Berlin,

Heidelberg.

Van Valen, L. 1973. A new evolutionary law. Evolutionary Theory 1: 1-30.

_______. 1976. Domains, deduction, the predictive method, and Darwin. Evolutionary

Theory 1: 231-245.

_______. 1985. Null hypotheses and prediction. Nature 314: 230.

Vrba, E. S. 1983. Macroevolutionary trends: new perspectives on the roles of adaptation

and incidental effect. Science 221: 387-389.

_______. 1989. Levels of selection and sorting with special reference to the species

level. Oxford Surveys in Evolutionary Biology 6: 111-168.

Vrba, E. S. and Gould, S. J. 1986. The hierarchical expansion of sorting and selection:

sorting and selection cannot be equated. Paleobiology 12: 217-228.

Wicken, J. S. 1984. On the increase in complexity in evolution. Pp. 89-112 in

Ho and Saunders 1984b.

Williams, G. C. 1992. Natural selection: domains, levels, and challenges. Oxford

University Press, Oxford.

Wilson, D. S. 1988. Holism and reductionism in evolutionary ecology. Oikos 53: 269-

273.

Page 32: On the Dual Nature of Chance in Evolutionary Biology and Paleobiology · 2018-07-03 · 1 On the dual nature of chance in evolutionary biology and paleobiology Gunther J. Eble Committee

31

Wimsatt, W. C. 1980. Randomness and perceived-randomness in evolutionary biology.

Synthese 43: 287-329.

_______. 1987. False models as means to truer theories. Pp. 23-55 in M. H. Nitecki

and A. Hoffman, eds. Neutral models in biology. Oxford University Press, New

York.

Wright, S. 1949. Adaptation and selection. Pp. 365-389 in G. L. Jepson, G.

G. Simpson, and E. Mayr, eds. Genetics, paleontology, and evolution. Princeton

University Press, Princeton, N. J.

_______. Comments on the preliminary working papers of Eden and Waddington. Pp.

117-120 in P. S. Moorehead and M. M. Kaplan, eds. Mathematical challenges to the

neo-Darwinian interpretation of evolution. Wistar Institutional Press, Philadelphia.

Page 33: On the Dual Nature of Chance in Evolutionary Biology and Paleobiology · 2018-07-03 · 1 On the dual nature of chance in evolutionary biology and paleobiology Gunther J. Eble Committee

32