1.1 Introduction
What are species? More exactly, what are separately evolving species lineages?
They might be feedback systems, displaying a distinctive degree or grade of metapopulation-level cohesion, which manifests and varies dynamically with the magnitude and frequency of feedback relations between several causal variables at population levels.
The next section briefly outlines that recent proposal (see Barker 2019a) â not to try to convince you it is true, but rather, to provide just one example of a new kind of species theory. This will set the stage for the chapterâs subsequent sections and its main aims, which involve next isolating a hard, long-standing problem for providing a theory of the species category (Section 1.3) and sorting species concepts that address that problem (Section 1.4). The chapter then uncovers instructive flaws in several views that imply we have either already solved that hard species problem or dissolved it altogether â so-called We Are Done views. The flaws are found in cause-focused species concepts, such as the biological species concept (BSC) (Section 1.5), and in non-cause-focused species concepts, such as the general lineage species concept (GLSC) and evolutionary species concept (EvSC). Other We Are Done views that are challenged include those stemming from Ereshefskyâs eliminative pluralism about the species category (Section 1.7), Mishlerâs pessimism about the category (Section 1.8), and Wilkinsâ phenomenalism about it (Section 1.9). More constructively, the chapter argues for a Revving Up view: rather than being done with the hard species problem, we are nearly ready to begin in earnest, as the feedback model will help exemplify.
1.2 Exemplifying a New Kind of Species Theory: Feedback Species
In a feedback system, processes loop or cycle. The values taken by variables over one time period feed back into the system, influencing the values taken by those same variables at later times. Put differently, output conditions turn around and serve as input conditions in further iterations of the same or similar pattern, as exhaust air produced from combustion in a turbo engine is fed back into the system to help (via interactions with other variables) amplify further combustion, thus increasing exhaust, and on and on until there is intervention or a dampening feedback mechanism is engaged. How about in an evolving species lineage? It is widely agreed that any species is a group of populations, a metapopulation (de Queiroz 2005). But, authors argue about which among a variety of processes or variables are most important for connecting and âholding togetherâ the populations in the group. Nearly everyone acknowledges that many variables often interact. But, those variables deemed most important are privileged over those with which they interact and are then referenced in differing definitions of species concepts to the exclusion of the other variables. Many have privileged gene flow processes in this way (e.g., Mayr 1942; 1963; Morjan and Rieseberg 2004; Bobay and Ochman 2017), others the sharing of selection regimes and adaptive zones (e.g., Van Valen 1976; Cohan 2002), and so on. Table 1.1 provides a partial list of variables that have been discussed. But fortunately, in recent years â including for both sexual and asexual populations â some experts have taken steps away from the privileging of just this or that small set of variables, making interactions between them central to their accounts of the nature of species (e.g., Templeton 1989; Boyd 1999; R.A. Wilson, Barker, and Brigandt 2007; Barker and Wilson 2010; Ellstrand 2014; Shapiro and Polz 2015; Novick and Doolittle 2021). The feedback model incorporates this but goes further, suggesting a long view on which interacting variables over one time period influence the values of these same and other variables at later time periods, which influences still later values of these variables at later time periods, and on and on over many periods, forming an evolving species system as the cycling feedback relations set it apart from other metapopulations.
TABLE 1.1 Provisional Running List of Causal Variables Entering Feedback Relations Variable label | Name of variable | Example works appealing to variables |
g | Gene flow | (Mayr 1963; Brooks and Wiley 1988; Morjan and Rieseberg 2004; Bobay and Ochman 2017) |
s | Shared selection regimes | (Ehrlich and Raven 1969; Van Valen 1976; Lande 1980; Mishler and Donoghue 1982; Templeton 1989; Cohan 2011; Shapiro and Polz 2015) |
h | Homeostatic developmental systems | (Ehrlich and Raven 1969; Mayr 1970; Wiley 1981) |
c | Colonisation | (Hellberg et al. 2002, 275â277) |
m | Mutation | (Mayr 1970; Hellberg et al. 2002, 275â277; Morjan and Rieseberg 2004) |
r | Genetic recombination | (Carson 1957; Mayr 1970) |
t | Trait similarities | (Barker 2019a) |
Source:Modified from Barker, M.J, Philosophy, Theory, and Practice in Biology, 11, 2019 |
For instance, gene flow between populations in a group over some time period, t1, distributes alleles between populations. That results at t2 in the populations having population-level trait frequencies closer to each other than to out-group populations. Subsequently, this helps ensure that the populations participate, over t3, in selection regimes in ways more like each other than out-group populations (e.g., Morjan and Rieseberg 2004). And in turn, that may help cause further gene flow at t4, and trait sharing at t5, and selection regime sharing at t6, and âŚ. The exact sequence neednât and surely wonât continue on with the same degree of regularity as in a turbo engine; a suite of variables may feed back into the system in diverse and changing ways over time, with varying degrees of influence or importance across cases. This suggests a relatively inclusive account of the species category (e.g., Ellstrand 2014; Novick and Doolittle 2021), on which diverse variables and relationships between them can be recognised as important across species groups.
However, the feedback model neednât be so inclusive as to imply that anything goes, or that species groups canât be distinguished from others. Part of the proposal is that relative to many non-species metapopulations, species groups feature distinguishing kinds of complex feedback regularity, even if this regularity is loose and liberal when compared instead with turbo engines. And, the model doesnât purchase inclusivity at the cost of vagueness, as weâll later see some other views of species do when they discuss the cohesion produced by species-forming and species-maintaining processes without comprehensively detailing the nature of such cohesion (e.g., Mayr 1963; Hull 1976; de Queiroz 1998; Wiley and Mayden 2000a; see Barker and Wilson 2010). On the feedback model, evolutionary cohesion is something that not only species but also other types of metapopulations can exhibit in different degrees or ways, and it can be given exacting but flexible mathematical descriptions that enable testable predictions. A basic central idea is that as feedback relations play out at the level of populations (within a group of them), this manifests metapopulation feedback cohesion, or M, at the higher level of the whole metapopulation (Barker 2019a). The values taken by recurring causal variables in the metapopulation feedback system can vary in magnitude, frequency, or both. When magnitude and frequency are both high across many variables, that manifests a high value for M at the level of the whole metapopulation system, perhaps indicating a species or even a sub-species. Low magnitudes and frequencies across many variables manifest low M, perhaps a genus or family. Intermediate values are of course possible. So, M at the metapopulation level varies dynamically with the magnitude and frequency of feedback relations between causal variables at population levels; evolving metapopulation lineages of many sorts may just be these dynamic feedback systems; lineages of the species type in particular may be those within a species-distinguishing range of M values.
Now, regarding prediction, consider a traditional view that has been widespread since the Modern Synthesis and will be part of this chapterâs focus. It proposes that within a variety of biological theories, the species category plays a more important or fundamental role than both more inclusive metapopulation categories (e.g., genus, family, or class) and less inclusive ones (e.g., sub-species, variety, or stirp) (see Wilkins 2018). As biologist Frank Zachos summarises it, âSpecies are the fundamental unit in many branches of biologyâ (Zachos 2015, 180). In combination with the feedback model, this could be developed to predict that weâll find a metapopulation level â the species level, presumably â at which the involved metapopulations clump in feedback variable space in a more distinctive or patterned and theoretically important way than at other metapopulation levels. Call this variable space M space (Barker 2019a). In it, each axis depicts the intensity (some combination of feedback and magnitude) of a particular variableâs feedback relations, a value whose measure is scaled between 0 and 1. For illustrative purposes, Figure 1.1 borrows from Peter-Godfrey Smithâs (2009) way of depicting such variable spaces (in a different, non-species context where feedback isnât central) and uses just three example variables as axes: gene flow (g), trait similarities between populations (t), and sharing of selection regimes (s). Suppose we plot a wide variety of metapopulations, from those hypothesised to be at or around the species level of inclusivity to many of those thought to be much more inclusive metapopulations. (The example and figure set aside less inclusive metapopulations.) The traditional view would then predict that we get a distinctive clustering of a small proportion of the metapopulations in some non-zero patch of the variable space, one of distinctive importance to biological theories. That would be the âspecies category patchâ, which may well have fuzzy boundaries but stand out to some quantifiable degree, with the involvement of some variables in feedback relations reinforcing the involvement of others in such relations. Along with this, the traditional view could predict that most of the remainder of the metapopulations (more ...