I realize that there are transgenerational selection effects, the classic example being the equilibrium pressure on sex ratios. The kin-selection example you gave is a good example of the more general notion of group-level selection, though due to its dependence on actual genetic relatedness it differs a bit from purely social forms of group selection, where a mere genetic propensity to join and behave appropriately is enough.

I certainly have more thinking to do about all of this; this conversation has been helpful in focusing the question of what “interests” are, and what can be said to have them.

If I may paraphrase, you appear to suggest that the interests of interest can be subsumed under the heading “maximizing fitness” which you qualify with “in the Darwinian sense.” I’m wondering whether you mean to contrast Darwinian with Fisherian notions of fitness. And, to clarify a bit why I’m wondering this, it was Fisher who proposed a transgenerational measure of fitness in the course of constructing a selectionist explanation of sex-ratios. In contrast, some people equate the ‘Darwinian fitness’ of an organism with the number of its direct offspring. On the one hand, you have a measure of fitness that tangles you in transgenerational timescales; on the other hand, you have a measure of fitness that is of limited use in selection theory.

And then there’s the notion of fitness at work in kin-selection models, where the locus of fitness is not the individual organism, but traits that can be dispersed quite widely within populations. While I appreciate the value of such models for elucidating the complexities of selection processes, I think this is not the most likely place to look for illumination about the content of an organism’s interests.

But, like I said, there are conceptual linkages between the notions of fitness and function, and I might well have mangled them in my own thinking about these things. In any case, we’ve only just begun the process of reordering our thinking in light of evolutionary theory, and it’s unlikely that anybody has it all “straight.”

I would say that the interests of an organism, very broadly speaking, are those things that tend to increase “fitness” in the Darwinian sense; often this corresponds closely to individual health and thriving, but often not. A salmon, for example, might have a longer life were it not to beat itself to death swimming upstream to spawn (just as governors might have longer political lives were they less concerned about spawning behavior also). And a mother will die trying to protect her offspring.

Evolution will optimize design to serve those interests. But there are different hierarchical levels at which this design “tuning” can happen; for example, social animals may exhibit optimzations in which their local, within-group success becomes secondary to behaviors that lift the overall fitness of the group — and it is fairly uncontroversial these days (though it wasn’t thirty years ago) that some selection occurs at the gene level also.

I agree with you that transgenerational timescales do not figure into any of this; evolution always reacts, of course, to the selection pressures of the present, and can’t look forward.

I think it’s probably right to view the interests of an organism as a sort of maximal realization of its natural functional capacities. Let’s call this “thriving” (actually, I think the more common word for this is ‘health’). It seems plausible to view thriving so understood to coincide with the interests of an organism. And let’s assume that the functional capacities the realization of which constitute an organism’s thriving can be individuated and explained in terms of the history of selection by which they evolved. So far, so good.

But now, consider how the process of natual selection that underwrites our attribution of natural functional capacities operates at the level of populations and on transgenerational timescales. Do the interests of an organism have such scope? I doubt it. After all, an organism’s realization of its natural functional capacities takes place in the here and now, and is what it is, regardless of what it might portend for future generations.

Right, that’s what I was getting at too. What was necessary for living things to have interests? There were conditions that had to be met or the organism’s line would end; only those organisms that met those conditions would stay in the game, and the better-equipped they were the better they did. An organism that was lackadaisical about food or sex might just get by, but an organism that had instantiated a ravenous “interest” in such things, along with tools and weapons for serving that interest would, as they say, eat the lazy one’s lunch.

When I mentioned Smolin’s universes as potentially “having interests”, it was a thought that had only just occurred to me at that moment, on the basis that if:

a) Natural selection operating on entities that replicate with heritable variation is indeed the only natural process we know of for creating entities with interests, and appears to have been what did the job in our case;

… and if, as Smolin suggests:

b) Universes themselves can be the products of natural selection;

then to be consistent one needs at least to ask the question: why are only some, but not all, of the things that can be designed by natural selection the kinds of things that can have interests?

But all that is going out on a very shaky limb; Smolin’s idea is very speculative indeed. Let me assure you that I am not trying to make the case that universes can have “interests”; all I was suggesting was that if universes, as Smolin proposes, really do exhibit replication with heritable variation, and are differentially selected in some way, then it becomes interesting to wonder why they can’t.

Again I think that we are talking past each other a bit here; I do want to be clear that I quite agree with you that there is nothing representational or goal-directed about natural selection itself.

But I do also think it is more than a metaphorical figure of speech to say that a snake has goals and interests; it is more than just instrumentalism, I think, to say that we can see that it really has these goals and interests even if it can’t.

I think that the snake, as a living, and therefore “designed”, creature, has goals and interests every bit as much as we do, with the only qualitatively important difference being our additional meta-representations of them (and of course our conscious awareness, but that is another matter altogether, and irrelevant to the having of goals and interests).

The evolutionary process is, in my view, the only persuasive account of how goals, and purposes, and intentionality could enter the world in the first place. But to be clear once again: this is very different from suggesting that the evolutionary process itself has goals or interests.

I think that maybe Deogowulf’s suggestion should be taken seriously. Dennett is very wishy-washy about what counts as a representation, a goal, etc. This is part and parcel of his studied ambiguity about whether these are just facons de parler or whether they refer to actual existents. When in his instrumentalist moods, he suggests it doesn’t matter. I, however, am clearly biased in thinking that science is about existents, the furnishings of this world. That’s why it’s important to me to employ language in a way that doesn’t tempt people to speak of the behavior of water in terms that, taken literally, suggest it has interests and goals. That’s why, even though I actually value metaphorical expression, I insist that in science, we must be able, in principle, to banish metaphor from what I’ve referred to as “positive theory.”

In other words, when it comes to “oughts”, what I say goes!

(Your “oughts” may vary.)

Of course, it wouldn’t surprise you to hear that I think Dennett — and you — are “getting out of the thicket” by redefining it as a clearing. (“A designed thing, then, is either a living thing or a part of a living thing, or the artifact of a living thing” — strikes me as very Dennettian.) As for being a materialist, I am not of any kind at all, as far as I can see, though no doubt, like most people, I have been rather hazy on the question. Still, if a man is to be a materialist, I believe he ought to be thoroughgoing one, with no inconsistencies or illicit appeals; the trouble is, with no inconsistencies or illicit appeals, can he believe he ought to be such?

I know he is a controversial fellow, but on this topic I think he’s quite right. I’ve been brooding about this stuff for decades, and this to me seems the clearest way out of the thicket. But then I think I am rather more of a thoroughgoing materialist than you are.

Stop reading Dennett!

Well, reverse-engineering an adaptation in terms of the interests of the organism to which it belongs could quite easily be part of a positive, predictive theory — for example one could predict how evolution will modify the trait in question as environmental factors change, etc.

I’m not clear what middle ground you have in mind between “explicit representation” of goal-states and “conscious meta-representation” of those states; from a design stance we can certainly see that the stalking behavior of a snake has as its goal the securing of prey, even if the snake has nothing on board above and beyond the wiring for seeking the prey.

To put that another way, the stalking behavior of a snake, and all the mechanics that have gone into making it function, are the result of a long and complex process of adaptation and modification — with successive iterations of the snake’s “design” or adaptation being adjusted, and this is the key point here, according to what does a better job of securing prey — in the way that falling water is not.

I think reverse engineering is a very useful strategy for developing hypotheses about the evolutionary origins of various traits. But that’s a matter of methodology, not positive theory.

I think we are more on the same page than it seems, and are experiencing the same difficulties in using these terms that so often complicate these discussions. I certainly did not mean to suggest that the process of evolution was “for” anything, or “signified” anything.

I disagree, though, about the uselesness of the idea of “interests” in understanding evolutionary products; quite to the contrary, often the best way to understand some adaptation is to do the usual thing that engineers do when reverse-engineering some artifact: to look at the feature in question and to ask how it serves the interests of the creature of which it is a part. Explicit representations of goal-states (that is, not only having the goals, but having conscious meta-representations of them) are a very late arrival, and just a fancy new feature built upon the underlying goals and interests themselves. Most living things just have the interests and act upon them, without ever having to “know” that they have them. A snake doesn’t have to be able to say to itself “I’m feeling hungry, and should go find some prey or I will be in trouble later on” — all it does is find prey when it’s hungry, and that works fine. Those explicit representations are, presumably, costly, and won’t be implemented if not needed.

Maybe you’re right, and “adaptation” is good enough. But it just doesn’t have the right “feel”, somehow — and think that is reflected in the fact that all these evolutionary biologists I cited above are certainly familiar with the word, but still use the word “design” instead.

As for ‘interests’, I think this term is so entangled with normative accounts of behavior that it is virtually useless in the context of evolutionary theory… until, that is, we enter into a discussion of how evolutionary processes have given rise to cognitive systems which employ explicit representations of goal-states. While I understand quite well what people are driving at when they speak of the interest of, say, genes in being replicated, I cannot view this as anything but poetic license. A rigorously formulated model of the processes involved in genic selection does not include any term or parameter that corresponds (not even as an analog) to the role of interests in contexts where the latter play an actual explanatory role. In other words, if it isn’t even a shorthand for some identifiable feature of actual explanatory models, talk of interests being served by the process of natural selection contributes zilch to our understanding of that process. I think what it does manage to do is to lull unprepared minds into a sort of complacency that is more likely than not to prevent them from appreciating how natural selection operates.

I’m not sure what it is that ‘adaptation’ falls short of. Since it’s used in evolutionary discourse to identify traits that have been shaped and/or maintained in existence under the influence of the process of natural selection, I should think it’s “reach” is precisely right.

We are a natural design, and do a fair amount of signification, though I might not be apprehending your meaning correctly.

I also think, given that I have insisted that the design process of evolution works with anything that replicates with variation and is subject to some sort of selection, that I must bite the bullet and say that I suppose I mustn’t rule out Smolin-style universes coming to have “interests” as a result of the evolutionary process (after all, that’s how we came to have ours) — though I have never thought about that, and can hardly imagine what form that might take.

Somehow “adaptation” falls a little short for me, but I agree it comes closest among existing terms.