Tautology 7: Conclusions 6 Sep 2009 So to finish, I will repeat the conclusion, and then make some comments on Fodor’s attack on “Darwinism”. Here is the complete series: The tautology problem Tautology 1a: corrections Tautology 1b: Butler Tautology 2: The problem arises Tautology 3: The problem spreads Tautology 4: What is a tautology? Tautology 5a: The issues Tautology 5b: The issues, continued Tautology 6: A resolution This will be a paper, and you will be able to download the ms version sometime later. I last said that the PNS itself is empirically empty, it forms a kind of schematic explanation: P1. PNS (+ ancillary principles, making up M) P2. Empirical data (O) P3. Interpretation Rules (I) ====== NS Prediction (or retrodiction) The schema provides us with what Aristotle might have thought of as a formal cause (aitos, or account); a set of explanatory structures into which we can place specific cases. It is not a true explanation until interpreted in a particular case, what Brandon refers to as “instantiating” the theory. Each part of the schema is substituted by some actual data or organisms, processes and parts. Then, it is an explanation. For example, while the same model may apply to a fungus or a fig, there is no physical property that they share uniquely. Without that, explanations by NS are merely promissory notes that there may be an account in terms of the ENS model. Given that until the model is physical and instantiated, any set of population dynamics might have turned out to be the one that best represented the particular case, it is hardly a tautology in practical terms. It may have been that the trait under investigation in that population is undergoing no selection at all. It may be drifting, or be hitchhiking on some trait with which it is linked that is undergoing selection. Even the various mathematical elaborations of the PNS, such as Price’s Equation, are tautologous in this sense. Okasha, for instance, says (2005): Though Price’s equation is really no more than a mathematical tautology, applicable to any evolving system, it has proved valuable for thinking about many evolutionary problems, particularly where multiple levels of selection are involved. This is exactly what I would have expected to find, given this account of the ENS as a schematic. In the case of multiple level selection theory, the tautology allows us to work out an interpretation in which the conditions of the equation, and the entities it posits with its variable classes, can be discovered in the empirical observations. In short, it tells us what to go looking for. Godfrey-Smith’s (2009) approach in which population is a Darwinian primitive is a similar case. We know that biological populations are the sine qua non for evolution; we think these things over here evolve, so what is the population (and by implication, the reproducers of that population)? The answer is not one that we bring to the particular case, necessarily, and it may turn out that common wisdom is not to be followed and that the reproducers are, in fact, the colonies or some other superorganismal entity. The schematic provides us with a benchmark to begin with, and a standard to measure our progress against. The interpretation is what we refine, not the theory as such.* Selection about fitness Finally, I would like to make some comments about a recent furore deriving from Jerry Fodor’s comments in “Against Darwinism” (Fodor 2008a; cf. Dennett 2008; Godfrey-Smith 2008; Sober 2008b; Fodor 2008b). Here, Fodor argues that because NS does not provide us with an account of intentionality, or “aboutness” (for example, why is it that selection has made the frog’s reflex to catch flies with its tongue about the flies, and not just any fast moving small black objects?) that the PNS cannot provide a prior story about selection “for”, as Sober calls it. Hence, he says, natural selection cannot provide an account of how selection operates. The way he frames it is complex, but it boils down, I think, to a tautology argument about what counts as fitness, and basically the loss of the “selection for” account (in the context of evolutionary adaptationism, which he misleadingly calls “Darwinism”†) means that the PNS is not able to give reason to think that we have whatever psychological traits we do (such as cheater detection modules) for adaptive reasons. This has two connections to our topic: one, that it is a form of tautology argument, and two, that it means if correct that any account of science being “about” the world (that is, representing the world truthfully or reliably) through a process of trial and error, like the PNS, fails. Since I want to argue that science is indeed a PNS-style process which results in there being representations “about” the world that are successful, Fodor’s attack has a dual import here. His argument is twofold: if NS is supposed to be about traits, then either (i) there must be intention in the ordinary sense; that is, there must be a selector such as Mother Nature, which is silly, or (ii) there are no laws of relative fitness (because of ceteris paribus considerations, etc.). Hence, NS is not an explanation for the fitness target. In our terms, we might recast Fodor more traditionally and have him say that there are no independent reasons to think that fitness is not a post hoc property, and hence it explains nothing. I think that the Brandonesque view I have been arguing for here overcomes Fodor’s rather gaudy objections. If the PNS is a schema, then it is a promissory note for an explanation; in and of itself it makes no empirical claims, of course. But when fleshed out, so to speak, it is a very good explanation. And since science, as an ENS-process (call it “Darwinism” if you really want to) has local explanations why this or that theory or model applies, the same thing is true there also, mutatis mutandis. But that doesn’t resolve the “intentionality” question, and the reason is that there is another conclusion to draw from an ENS account of anything: intentionality is not a real thing either. A theory that does duty, whether it is properly about the domain or not, and a trait that increases reproductive value, whether it is about the environmental property or not, will, necessarily, ceteris paribus, increase in a suitable population over a suitable amount of time. Nothing whatsoever is added to the account by making frog actions “about” flies, so long as those actions increase the reproductive value of whatever biological substrate is passed on through inheritance. In an ecology where BB pellets outnumber flies, those actions may be fitness decreasing; so what? By privileging the intentional aspects of models and of behaviours and traits, we have committed a fallacy, of reifying abstractions. It may serve us to think that a certain word or representation is about something else, or that a certain trait adapts to an aspect of the ecology, but really, selection is a sorting process, and has no intentionality at all, and so intentionality is, I think, a useful fiction. It is selection that is primary here, in general terms. Each case, like each unhappy family, goes its own way in realizing that, or not. Notes * Clearly, theories that are persistently inadequate to the observations in a domain will be refined. This is merely meant to apply in a single ordinary case of scientific explanation. † Few terms have the total lack of clarity that the term “Darwinism” does. It should never be used by anyone who wishes to make their target or purpose clear. References Dennett, Daniel. 2008. Fun and Games in Fantasyland. Mind & Language 23 (1):25-31. Fodor, Jerry. 2008. Against Darwinism. Mind & Language 23 (1):1-24. Fodor, Jerry. 2008. Replies. Mind & Language 23 (1):50-57. Godfrey-Smith, Peter. 2008. Explanation in Evolutionary Biology: Comments on Fodor. Mind & Language 23 (1):32-41. Godfrey-Smith, Peter. 2009. Darwinian populations and natural selection. Oxford: Oxford University Press. Okasha, Samir. 2005. Maynard Smith on the Levels of Selection Question. Biology and Philosophy 20 (5):989-1010. Sober, Elliott. 2008. Fodor’s Bubbe Meise Against Darwinism. Mind & Language 23 (1):42-49. Epistemology Evolution General Science Philosophy Science Truisms
Australian stuff Why anti science? 23 Nov 2013 Over the past few decades there has been an increasingly negative attitude by governments, pundits, religiosi and faux philosophers against science. We have seen an increase in denialism about climate change (one of the most well supported scientific models of the day), vaccination, evolution, medical research in general, and the… Read More
Biology Rise of the Planet of the Moralists 2: chains and trees 18 Oct 201122 Jun 2018 Rise of the Planet of the Moralists Series1: Introduction2: Chains and Trees 3: Clades and grades4: Predicting traits5: Social dominance and power Very famously, Darwin came up with the idea of the evolutionary tree. What is not often realized is that it is the tree that is more uniquely Darwin’s than natural selection, for which… Read More
Evolution Sociobiology 2: Theoretical foundations 18 Nov 2007 Wilson and Wilson begin by reviewing the reasons why sociobiology of the 1970s was rejected. They focus on the arguments against group selection. Read More
<i>but really, selection is a sorting process,</i> Hmmm, I thought it wasn't even a process or mechanism. Eliminative or negative selection might be, though. Basic consciousness or subjective awareness is not just a process, otherwise why-am-I-me could be easily explained in the context of an objective physical model with no preferred locations. Perception however, could be a process….
Good roundup — looks like a lot of philosophers have been hung up on “aboutness” lately as some sort of tangible, mystical, almost supernatural entity. I think it’s Searle’s main mistake in the Chinese Room but until now I haven’t realised that it might possibly relate to Fodor on “Darwinism” — since both appear to rely on our intuitions of some mysterious semantic content that have little to do with the processes (NS/consciousness) at hand. And the main problem is how hard to be clear on this subject! This very comment is probably ambiguous/unclear as well and would need to become a full paper to be a “proper” comment. Sigh.
Thanks. I’m coming to a full-blown eliminationism of matters non-physical. I’m already denying the extra-mental reality of information, intentionality, functions, and rejecting hylomorphism of all kinds. At this rate I’ll have to go live in Quine’s desert.
There’s a self-referential or circular problem here – it is impossible to think of any thought that is not “about” something, your comments on intentionality included 😉
Yes, a good summary… but, of course, I have a quibble. Fodor may well be “hung up” about the use of intentional idioms in discussions of natural selection; idioms that John (rightly) thinks can be avoided without damage to the theory of natural selection. But remember that Fodor hails from the cognitive side of the sciences, where intentionality cannot be dismissed with a simple “So what?”. Selectionist models applied to cognitive practices (like scientific endeavors) will have to accommodate full-blooded intentionality.
Well actually I do think that we can dispense with intentionality in cognitive science. Sure, we like to think that our thoughts are about this or that, but they are just what they are, and they achieve what they achieve. “Aboutness” is either something very obvious or it is an occult property.
but really, selection is a sorting process, Hmmm, I thought it wasn’t even a process or mechanism. Eliminative or negative selection might be, though. Basic consciousness or subjective awareness is not just a process, otherwise why-am-I-me could be easily explained in the context of an objective physical model with no preferred locations. Perception however, could be a process.
Sorting processes are processes that can be described with a sorting algorithm. Selection is a kind of sorting process, and so can be described by an algorithm.
In tautology 3 you seemed to the question the idea of selection (real world) as a process or mechanism, no doubt due to the “tautology” and the difficulty in separating fitness and selection. But maybe you were simply characterizing the position of others.
I was a bit careless, addressing intentionality with its connotations of “aboutness,” where Fodor emphasizes intensionality, serving primarily to contrast with extensionailty — distinct but related matters. Be that as it may, “aboutness” might be both obvious and occult; similar in that respect to gravity.
Nice work. It seems like you’re in accordance with Dennett about intentionality’s issues, am I wrong? Or are you suggesting some strong eliminativism, like the Churchlands do? Anyway, I’m also coming to deny stuff like mental representations in my ontology, but sometimes they are seductive (they may help when you’re trying to make psychology work well with semantics) and in the end of the day teleosemantics begins to sound better and better.
I’m often with Dennett on these issues, and occasionally the Churchlands. Here I think representations exist (but not in the head so much as between heads), but they have no magic noetic denotation caloric.
Sorry, I posted that before I had finished. As to teleosemantics, that relies, I think, on the notion that there is selection for traits. Here I agree with Fodor; there is no way selection can be “for” anything on its own. We have to isolate the traits based on our own reconstruction, and so selection for is not mind-independent; i.e., it’s not in the world. This puts me at odds with Dennett/Sober.
Please reflect on the difference between intentionality and intensionality — “selection for” is generally understood to involve the latter, but not the former.
Bob, I think you have it backwards. According, for example, to the Stanford article on Intentionality (section 9) selection is supposed to be an intentionality provider. Dennett says the same thing. There are other examples easily found with a quick Google: here, for example. But I would appreciate you explaining why you think this – there may be something I have missed.
John – I don’t think I’ve got it backwards… As the SEP article notes, “extensionality and intensionality are logical features of words and sentences,” whereas “intentionality is the power of minds to be about, to represent, or to stand for, things, properties and states of affairs.” The “selection for” locution exhibits the logical features of interest, most notably, failure of substitutivity of co-extensive terms — and this is what has Fodor so exercised. He is quite clear that he doesn’t think natural selection exhibits intentionality or “aboutness.”
Yes, he is, and that is why I have accepted his tine of the fork, and simply denied that intentionality is required, contrary to Dennett’s, Millikan’s, Neander’s and others’ view that the teleosemantics are given intentionality by selection for. Simply because selection fails to deal with intensional substitution, I think that meaning has other foundations (or rather, few foundations at all beyond use) but there is no doubt that intensionality can be present without there being an intentionality. I think that you are over-specifying things if you (like Fodor) think that intentionality can only apply to representations – in cognitive science (such as Searle’s argument, as Michael noted above, it refers to the aboutness of belief states, yes, but it also applies to anything that is about some referent, as Fodor argues with the flies example). I think two things, as clearly as I can express them at 2am: 1. Natural selection does not give us intentionality (as Fodor says) 2. The meaning/use of many words and sentences, their truth value and denotation, are given by a process of natural selection (as Dennett, Millikan, et al say). I conclude that expressions do not have intentionality in any real sense. We say they do, but that’s just disquotation; we are no less in the semantic prison after than before. We are being misled by our privileging words and thoughts over the world. Reread the Dennett link. He sets it out (but thinks that teleosemantics works) the way I have done.
John – I, too, think that something like the “aboutness” taken to be characteristic of intentionality is realized by bits of the world independently of minds and, more pertinent to this discusion, independently of any processes of natural selection. That’s because I see information as what Dretske calls an “objective commodity.” As I view matters, it’s in virtue of such objectivity that informational relations are “accessible” to strictly objective causal processess that instantiate natural selective processes (that’s a mouthful). So I’m not suggesting that selection “for” is the source of intentional “aboutness”. I would also argue, somewhat idiosyncratically, that it is only when selection “for” informational content occurs that structures bearing the information in question acquire representational functions. So I’m not suggesting that selection “for” is the source of intentional “aboutness”. In this I part ways with most teleosemanticists, though I embrace much else in their program. But I think this is largely beside the point for Fodor’s argument. If I understand his argument (of course, I might not…), it’s about the lack of fit between the intensionality of “selection for” and a particular view about scientific laws and their role in explanation. It’s his view about laws and explanation, not his grasp of evolutionary theory, that I think needs to be addressed.
I conclude that expressions do not have intentionality in any real sense. We say they do, but that’s just disquotation; we are no less in the semantic prison after than before. We are being misled by our privileging words and thoughts over the world. And this is where we run into a problem with eliminationism. There’s plenty of intentionality, ‘aboutness’, and meaning in your statement, and yet you would say that all other forms of intention and meaning are misleading – except yours. Isn’t this at least somewhat circular and self-refuting? (as per Boghossian)
I’m not saying that there isn’t “aboutness” in the semantics. I’m saying that isn’t a physical or metaphysically independent relation. Nothing makes my statement “George is a dog” about George or dogs except that in this language system, social context and historical nexus, such things are referred to in the language. It might be that I think that a cauliflower is George, but that is a failure internal to the language game. If the language community reliably connects its terms to biological individuals given the social appellation “George”, and these individuals are of the class “dog” then there is aboutness; but the aboutness is not some real nonlinguistic relation.
but the aboutness is not some real nonlinguistic relation. One must ask then, if it is possible to discuss anything independently of lingustic relations or aboutness. If not, then one could make the case that you presume a metaphysical reality that you cannot describe. The buck has to stop somewhere.
Maybe I got things wrong, but this is how I see the relation between intensionality and intentionality: there’s no intensionality without intentionality, but there can be intentionality without intensionality. Intensions are semantic properties normally identified with ‘definition-like’ logical strcutures (or with fregean senses, concepts, etc.), which means they have epistemic sensitivity and hence psychological sensitivity, therefore you can’t argue for the existence of intensional properties without commiting yorself to some sort of intentionality. And that there can be intentionality (or meaning) without intensionality is one of the conclusions that can be extracted from the direct reference theories. I agree with John that selection fails to deal with intensional substitution. I don’t even see how intenSional properties could be instantiated by selection. Intensions are invisible to natural selection.
If Dretske is right, and information is an “objective commodity,” then there’s a sort of objective “aboutness” even if there are no “intentional systems” of the sort cognitive scientists study. One bit of the world can carry information about another bit of the world. For example, tree rings can carry information about changing temperatures; or an optic array can carry information about the relative positions of various reflective objects. My suggestion is that if this was not so, it would be impossible for intentional systems to evolve.
I’m not at all sure that Dretske is right about information. Certainly, we have evolved information detection and processing systems in the animal kingdom, but that hardly implies that information is an objective commodity. What counts as information depends crucially on the Umwelt of the organism, and this is phylogenetically distributed, so I can’t conclude that information is an objective metric.
For example, tree rings can carry information about changing temperatures;… Do they? Using theories on growth rates of trees and environmental factors we turn measurments of tree rings into information about climate changes during the life of the tree. The information is not in the tree rings but is created by the dendro-scientists.
Thony C – So, dendro-dudes didn’t discover the informational relations that obtain between tree rings and ambient temperate? Is there then no fact of the matter about those relations?
John says, “What counts as information depends crucially on the Umwelt of the organism.” I’d like to see how this gets worked out in a story about the evolution of detector or processing systems. What is it that’s getting detected and/or processed? I say the alternative view can’t simply be ignored; namely, of the information present, the part that is accessible for processing by an organism depends crucially on the Umwelt of the organism. This seems to be almost a truism. It is good to bear in mind that Dretske’s account of information is driven by his epistemic program, and this leads him to emphasize abstract features of informational relations. For an equally objective account of information addressing an ecological problematic, see anything by J.J. Gibson.
Take the patterns on flowers that are visible only in bee purple UV. They are signals to that bee, and not to us. For us, they are simply white flowers. What signal do we get from white flowers? Maybe that it is spring. The bee gets a whole other signal (come get pollen here). Is there a signal objectively in the world? I don’t think so. Instead there is a perfectly physical causal story to apply in terms of coevolution, neurological and ultimately physicochemical mechanisms, and contingent conditions. We choose to give this an information spin, because the math and metaphor suits our own cognitive propensities and those of the scientific process. But there is no objective (i.e., not subjective, or observer dependent) property “out there” that answer only to “information” however conceived. This also applies to Gibson’s ecological psychology.
John – That something is a signal to bees but not to us fits perfectly with the objectivist view I sketched. And the availability of “a perfectly physical causal story” does nothing to undermine the objectivity of informational relations — rather, it’s the basis of that objectivity. The suggestion that something should answer _only_ to the “information” appellation is a red herring. Nothing in this world, so far as I know, answers to _only_ one appellation.
This little article by Massimo Pigliucci on “Why There Can’t be a Theory of Everything” might be relevant to this discussion, and to anyone who maintains that there is some viewpoint, inference, or “ism” that ultimately describes everything: http://www.livescience.com/strangenews/081024-sb-theory-of-everything.html
Thony C says, “We construct information out of those causal relation the relations themselves are not information.” I don’t know what this process of constructing information amounts to. What does it mean to say that it’s constructed from causal relations? Also, one wants to say, “Of course, it’s the content that constitutes information, not the causal relations, even if they play a crucial role in determining what the content in question is.
It’s vital to not conflate different senses of “information”: signal, evidence and functionality are all distinct meanings. When two birds sing to each other, that’s a signal. There are information processing systems in each that are able to “transmit and receive” because they share the same protocols (via biological and cultural evolution). But we and the tree do not share such protocols. The tree rings are “information” to us in a much different sense: we have knowledge that makes the tree rings evidence of annual growth. This sense of information, sometimes called Fisher information after R. A. Fisher, is basically the likelihood that the measurement is correct (imagine a season where a false summer makes the tree have a second burst of growth, or a bad summer, no growth at all). What we have are causal processes leaving evidence. That becomes information so long as we are able to read it (or otherwise it’s just a pretty pattern, signifying nothing). Hence it depends crucially upon our assay abilities. That is what I think Thony meant (and what I would mean). “Information” is a homonym.
It is indeed what I meant and indeed still mean. Wonderfully expressed Mr Wilkins; you should become a philosopher, you know!
I don’t think I’ve been conflating different senses of information, but it’s still good avice to avoid doing so. I’ve noted a couple times that the way I’m using the term has affinities to the uses of Dretske and Gibson — and while I recognize difficult problems in both of their accounts, I don’t think those problems would be mitigated in the least by dropping the assumption that information is objective. I don’t know how to interpret ‘information’ to make sense of the idea that evidence becomes information as a consequence of our being able to read it. Was it evidence independently of our being able to read it, or are you just shifting the problem? (‘Evidence’ seems a prime candidate for explication in information theoretic terms; ie., and crudely stated, it’s in virtue of its informational content that something can serve as evidence. At one point, you even say that evidence is one of several senses of ‘information.’ But then the idea that evidence becomes information is quite confused.)
The release of Fodor’s book has got me thinking about all this again, and your summary here of the tautology problem is invaluable. It’s noteworthy that Lewontin, in his non-tautological description of NS, nowhere mentions fitness or adaptedness; only heritable variation. This is somewhat heretical. I also notice he says that his description allows us to make predictions, but this only seems to follow if the variations are known in advance, which is giving the game away. Neo-Darwinism does not, so far as I know, suggest we can anticipate random point mutation, or (given the facts of pleiotropy and genomic plasticity) these mutations’ exact effect on the phenotype. I’m also interested in Popper’s recantation, where he says “not everything that evolves is useful… though many things are.” This gets away from the tautology problem, but at a deep epistemological price. If some traits are neutral (as they surely are) we are back to common sense to define what is an “adaptation” and what is not. Spandrels are one black box, and Orgel’s law is another. If evolution is smarter than you, then what is apparent about a morphological development may not be what is pertinent. (This goes to the Umwelt/Ecological psychology problem too. We can only ask the questions about nature that make sense to us to ask–[itself a tautology].) Gould’s reply to Bethell isn’t much of a “refutation” (as one of your earlier commenters mentioned). Gould says that NS is not a tautology if it can be shown to have “independent criteria.” And what are those criteria? “Superior design in changed environments.” This is just a restatement of “fitness,” and it is just as context-specific. What makes the design “superior”? The differential survival. And so on. This definition can be made consistent with the Spandrels argument only by re-inserting intensionality–second-guessing–which is Fodor’s argument. Either (1) NS is directly linked to differential survival, in which case nothing new has been added to our understanding; or (2) we resile to just-so stories about how the leopard got its spots. Neo-Darwinians have been able to skirt this problem by attenuating between the extensional and intensional definitions. Drawing a circle around them both at once may help to get past the error.
If some traits are neutral (as they surely are) we are back to common sense to define what is an “adaptation” and what is not. …just as risk assessment is difficult and counterintuitive, so is the concept that randomness or chance can result in a “directed” change. And so tiny risks are inflated out of proportion, and the unpredictable components of evolution are discounted.