This is possibly one of the reasons analytic philosophy is such a graveyard of ideas: ... Propositions - yuck. — StreetlightX

Eh. The single most famous paper in the modern era of analytic philosophy, "Two Dogmas of Empiricism," is a full-out assault on the idea of "the proposition."

Ceteris paribus? It's "in principle" after all.

You could head for something like: given a set of beliefs, belief P is the belief which, if added to the set, eventuates in action A.

Or: belief P explains action A if P and only P is a member of every set of beliefs that explains A.

I can think of lots of models to test drive.

I tried to defend the notion that to believe something is to act as if it is true. It didn't work — Banno

As you and recently @Janus have noted, in a propositional account of belief one argues not that belief must be expressed in fact, but that it must be expressible in principle.

Why not follow a similar line here? You could argue that beliefs are things that could be acted upon. Or as you put it, it must be possible to behave as if the belief were true. (And for free you get a way of distinguishing beliefs from each other.)

They're closely related, of course. One old test to show that A presupposes B is if both A and ~A entail B. (Present king of France and all that.) That doesn't seem to be the case here, so the modern technical sense of presupposition doesn't fit.

Both sometimes seem to run backward "causally", in the sense that if A entails B, that could mean B is a condition for having A, and that's not too far from how "presuppose" is ordinarily used (leaving aside the modern technical sense in philosophy and linguistics).

Now I can't tell if Schopenhauer is saying B is a condition for A, perhaps something you must have first, in temporal order, maybe something you could have by itself without moving on to A; or just that B is part of the meaning of A, a necessary part, that there is no kind of A that isn't B.

There are lots of different sorts of things that can be expressed by saying "A entails B", "A implies B", "A only if B", "No A without B". Some of those things ordinary folks think of as presupposition, some philosophers and linguists think of as implication, presupposition, or implicature (although that one I'm not seeing here).

Maybe it's crystal clear in the German.

(Sorry if this is just stuff you already know.)

Is there any reason to think that a plant in need of water might refuse it?

I want to start by saying I love the description of yours I quoted above:

An organism must be able to both persist and to adapt. In the long run, it must be stably centred or balanced - hence homeostasis. In the short run, it must be able to adjust that general balance in locally useful ways. — apokrisis

I think that really captures something nicely: I picture the difference between a boulder rolling down a hill at a fallen tree and a deer running down a hill at a fallen tree: the boulder just plows into it, but the deer leaps gracefully over it.

Think of the different ways of drinking water:
1. A towel placed on a spill will "drink" the water, but this is a purely mechanical effect. The towel has no agency here; it's not "doing" anything. Absorbing spills is the function of the towel, the reason for which it exists, the use to which it is put, and that's one sense in which the word "purpose" is used. But the towel absorbing water is not intentional or purposeful on the part of the towel. And the water is not of benefit to the towel.
2. A deer seeks out water periodically because water is of benefit to the deer. Is this intentional, purposeful behavior? Well, the actions the deer takes to get water might be, might not -- I don't know much about the inner life of deer. Trees send roots into the soil because water is of benefit to trees, and I have no reason to think trees have inner lives. The mechanical process by which roots take up water is probably not much different from the towel's.
3. We drink water much as deer and trees and towels, but we can also choose not to, for any number of reasons. When we do so, we have agency, our action is intentional and purposeful, but it is not our purposefulness that makes water have benefit for us.

One of the things I learned from our last conversation about purpose is that an action can be construed as serving any number of purposes, and that the goals actions serve can also in turn be construed as means for reaching other goals. Purpose is a never-ending hall of mirrors.

What I want to do here with "benefit" is cut that off: water is of benefit to the tree as a tree, for it to persist, as you said, as a tree, doing whatever it is trees do, whatever it would make sense for trees to think and talk about if they thought and talked. The towel is a particular, but it is not an individual, and so nothing can be of benefit to the towel. Purpose in the sense of function, you can find all over the place; purpose in the other sense, I think only makes sense for us and whatever critters have inner lives enough like ours.

To your specific questions:

But why wouldn't there be a direct connection between purpose and benefit? — apokrisis

Obviously for critters like us there often is, but it's optional.

What would a benefit-less purpose even be? — apokrisis

It should be clear by now that I mostly mean benefit as benefit-to-the-organism whose behavior we're looking at. So I'm happy allowing humans, for instance, to have purposes that do not benefit them. Guy saves a guy at the cost of his own life -- I don't need to concoct some benefit to him to "explain" that. It's just what he chose to do.

What would a benefit be except that it served some purpose? — apokrisis

See that's where I think there's just too much slippage between the senses of "purpose". Water benefits trees, it has a use within a tree, serves a purpose -- but we're just talking functionally here. There's nothing like intentional behavior in the water or in the tree, so I don't see any purpose in that sense.

Despite teleology being deemed erroneous by the prevailing materialist metaphysics of the day, you’ll notice that in our mode of thinking teleology will be intrinsic to both aspects you address: something being done for the purpose of some given X; e.g. “using fuel” for the purpose of (i.e., because of the need of) “creating energy”, or “reproducing” for the purpose of (as one example) “preserving one’s own identity”. In both examples, the latter is the telos to the former activity. — javra

Do you not see anticipation in photosynthesis, seed production, and growing in general? How can anyone deny that these are goal oriented, purposeful? — Metaphysician Undercover

An organism must be able to both persist and to adapt. In the long run, it must be stably centred or balanced - hence homeostasis. In the short run, it must be able to adjust that general balance in locally useful ways. — apokrisis

It can be convenient and useful to refer to trees as acting in accord with purpose, or motive, or telos, but these accounts are simply abstract fictions, there being nothing in the tree purpose or motive or telos occurs. — tim wood

The above are mostly intended to be representative quotes. I think the trouble in this discussion comes from trying to fit the square peg of purpose into the round hole of benefit.

I don't disagree — Sapientia

What trickery is this?!

Okay, but there's a natural follow-up to what I wrote, which is to tell a story something like this: there is progress in rationality by not filtering what your conviction engine is exposed to, and further progress in deliberately seeking out disconfirmation of your working theory. I like that story. And it gives you a ready way of ranking norms: if you must ignore evidence you don't like, I count you as less rational.

Your point that there is action taken that only indirectly affects conviction stands, because we can try to follow a course of action without a chosen outcome.

That's all quite idealized, of course, as recent issues in social science research show. So then we're back to questions about how the course of action is chosen.

Some quotes I use as maxims:

Kids need your love most when they deserve it least. — Erma Bombeck

Everyone has their reasons.

That one I got from Orson Welles, explaining Touch of Evil , and he said he got it from Renoir.

Suppose I now said that in the next paragraph I'm going to present an argument against your position and in favor of mine. You could choose not to read it, and avoid exposing your inner conviction engine to something that might, without your control, move you over to my side. There are people who do exactly this.

"Act as if ye had faith, and faith will be given to you" is the old Catholic saying. Even if you cannot choose to hold a belief, if you can manipulate the circumstances of belief formation, then there's something choice-like going on there. People do talk themselves into beliefs, and arguably reasoning is exactly such a process.

We can of course just deny that there's choice in any of these behaviors, but that's a whole 'nother deal.

So it definitely wasn't his viewpoints. — Michael

It was not because of his viewpoints — StreetlightX

That's what I meant.

The started reasons for the ban, combined with intransigence, I find satisfactory. @Baden's initial comment

unreasonable antagonism towards science was a part of it — Baden

was a little disturbing, but it was clear he didn't have first-hand knowledge.

I don't think it was his viewpoints that actually got him banned — fdrake

I for one would be happier if someone said it definitely wasn't his viewpoints, all of which i disagreed with.

the presumed choice still seems to appear out of thin air without explanation — Sapientia

Just going off the the idea that "should" here is normative, that you should conclude such-and-such in the same sense that you should respect your elders or that you should eat your vegetables. Norms are usually something you get to choose to conform to or not, and there are consequences either way.

If my reasoning doesn't match theirs, then that's that. What's the choice supposed to be? The only choice that I see would be the choice to conform disingenuously or stand by my reasoning. That's no choice at all for someone like me. — Sapientia

This is an excellent point. Most of the time, we no more choose to reason as we do than we choose to speak our native language as we do. We've learned, we've been trained, we've developed habits, and so on.

But I've been in the position, on this very forum, of arguing that if you say A then you should say B, and it sure feels like I'm saying that this is the right way to reason, that I'm invoking a real norm.

The conundrum stands...

I'm not sure your example is the same though. Yours I hear as a subjunctive, acknowledging gravity as an hypothesis.

I took my "should" as normative: if you reason the way we do around here, this is what you'll conclude. Not an expectation in the predictive sense, but in the "it's mandatory" sense. (Every one of these words tries to be ambiguous in the same way...)

I thought that likely, as a matter of fact, this being a phone, and me hating writing on it, but the flag button doesn't show up at all until it's posted... Not gonna worry about it.

Cool. But how could it get flagged without ever being posted?

It was immediate.

You can attempt to hold two opposites in your mind as true, or at least equally uncertain, but at the end of the day, you must act as if one is true. — AlmostOutlier

I think this is exactly what we do all the time, except without the "equally" bit. It's highly likely the ground you walk on will support your weight, so you walk on it as if it will, as if you know it will. But you know no such thing, and the occasional sinkhole makes this point forcefully.

The analogy I always reach for is wagering: you may figure the Cubs have a 62% chance of winning tonight, but you have to bet as if they will or won't. Confidence may fall along a gradient, but action is binary: you can kinda think the Cubs will win tonight, but you can't kinda bet.

The notion of picking beliefs strikes me as oxymoronic and disingenuous, like picking where you were born, your age, or who you'll bump into today. — Sapientia

I'm very sympathetic to this view. It does leave me with a bit of a puzzle though. If we do not choose our beliefs, should we say they are caused? In some scenarios that seems okay; perception seems like this.

But what about reasoning? Suppose A being the case makes it that B is. Then we would customarily say that believing A is a reason for believing B, that if you think A likely, for instance, then you should think B likely.

What do we mean by "should" there? That suggests you could choose to reason this way or not. Does that leave us saying you cannot choose simpliciter to believe B, but you can choose to conclude B by reasoning? Would you then be choosing to cause yourself to believe that B?

a strange time zone — Bitter Crank

Saw what you did there.

Are they indoctrinated (in some negative sense) to an individual through repetition and training? Would this be the primary function of 'schools' and 'education'? — Posty McPostface

Is teaching my daughter how to use a hammer indoctrinating her?

That's a nice summary. Chance and necessity make a nice pair of terms in which to explain everything, but I would imagine you could tell a similar story with other pairs (or mores) of fundamental somethings. They all make me uncomfortable, but that's my problem.

Pierce is interesting, on the one hand, for giving Chance a seat on the dais, right? And then you realize that if we are honest, and serious, about our model-building, it too will have a place for chance, at least as whatever it is our theory doesn't account for, the noise accompanying the precious signal. Now we get to match up the epistemic and ontological, as you've noted, and we can be indifferent to questions like, "Is chance real?" We can posit it, or not, but it will always be in the model either way. And this would be Peirce's pragmatism, yes?

Oddly, this matching up makes me even more uncomfortable than the Big Theories do on their own. If the big theories already seem to hang in the air (the way a brick doesn't) on the buoyancy of their own internal coherence, this version seems more like jumping and forgetting to hit the ground.

Again, my problem. If I've understood, what most convinces you you're on the right track is what gives me the willies, which is curious. As always, I appreciate your patience with my questions.

Can we identify a case where the " The whole is LESS than the sum of the parts?" — malcolm

The United States Congress.

Edit: possibly also the Los Angeles Dodgers.

The "map" is of the very fact that accidents accumulate to form the regularity of habits — apokrisis

This pattern I like very much, and I'm totally on board there.

But since we're talking metaphysics, do you have any qualms about the word "fact" here? What kind of fact? Are we forced to call such accumulation itself either accidental or necessary?

I would like to think this is where pragmatism slots into the story, but I don't quite get it. (Metaphysics just not my thing, as you know all too well.)

Sure history is full of accidents. But if these accidents can accumulate, then they become the constraints that act in the present to limit the accidents of the future.

They are no longer accidents once they become part of the constraints that prevail — apokrisis

That makes nice sense. Yesterday's chance is today's necessity. I understood your project to be pushing back or outward to ever greater generality, to the "purely" necessary. I guess if that's only an ideal, you'll be mapping the ossified accidental just like the rest of us. I suppose that's the sense of mapping "from the inside", as you put it.

So you are simply attempting to make an analogy the worst possible by abusing it in the worst way you can imagine. — apokrisis

Really not. Since what is essential depends on the purpose of the map, I couldn't see what a truly generic map could be. Since what today is a constraint might not have been, I couldn't see what a map of only the necessary would be. Your response helps. I still don't quite get the big picture, but I'm good for now.

Carry on.

My thought here was that the usefulness of a map is showing you what roads happen actually to exist connecting features you're interested in that also happen to exist, and it shows where the features and roads actually happen to be. You could abstract away location, distance, and so on, and just show the connections -- but this town and that city and the road that connects them are still matters of accidental history.

Constraints would only show you what connections could exist, where they could be, etc. We need to know which ones actually obtain.

Granted some features are considered essential to a map, in the sense that they're included when others aren't or needn't be, but it seemed to me those included features are still historical and accidental -- this town might not exist, there might not be a road between these two, etc.

It is just like real maps - the kind you use to get around. — apokrisis

This is just a metaphor, so whatever, but there are other kinds of maps.

the simplest map just tells you where are the obstacles, where are the paths — apokrisis

And that might be simplest for the kind of map you're talking about. But your idea seems to be there is a purest sort of map, a perfectly general and generic map, and that's just goofy, since what you choose to include in your map is obviously driven by your purpose in drawing it. The stuff you talk about leaving out is the stuff other people want.

In regards to the contrast between lived life and metaphysical maps, a map is created by abstracting away the accidental to arrive at the necessary.

So actual life is rich because it it rich with a history of accidents, fluctuations, contingencies and particulars. — apokrisis

And right there -- if you want your analogy to be to "getting around" maps rather than some other kind, you want the current state of accidental history. Is that bridge still up? Does this surface street go under the new freeway or just dead end there?

In fact I can't think of any kind of map that isn't based on selecting certain accidental states of affairs to mark and the rest to ignore. There's never any essential/accidental distinguishing such as you describe.

Sorry -- this just seems like the worst analogy for what you're after.

Now, see, that sounds largely reasonable, in the way that pragmatism always does, if a little empty. What I'm struggling to get across is the oddity of your position, and it's probably just the Peirce thing.

On the one hand, there is this sort of messianic quality to your system. From earlier in the post I responded to:

I agreed that modelling is modelling. But then the larger Peircean story is that modelling constructs its own world. And so the actualised wholeness is itself an emergent from the core semiotic process that is the engine producing any reality. — apokrisis

That doesn't look like the sort of instrumentalist version of pragmatism that this does:

Pragmatism says "wrongness" is to be expected. The question then becomes whether the wrongness observed as the general is advanced to explain the particular is a case of signal or noise. Is there something significant not being explained? Or are all the inevitable exceptions to the rule just meaningless noise? — apokrisis

The procedure by which we model our world is the same procedure the world is using to be.* Where is pragmatism here? Is reality also getting it wrong the way we do? In which case, the noise in our models is also signal, and what sense now can be given to "wrong"? If you put "wrongness" in scare-quotes, is this still pragmatism?


* What's more, projecting this structure onto nature was the "leap of faith" that was said to work. Why does it all look a bit like Anselm's ontological argument?

I've said nature itself is irreducibly telic. There is always finality or a goal in play. — apokrisis

And that's just not enough to start talking about what "works".

Here's the last part of the quote I started with:

we are projecting a view of ourselves as rational beings on to the apparent rational structure we see in the world. That is quite a leap of faith.

However - pragmatism again - suck it and see. Leaps of faith become justified to the extent they appear to work. — apokrisis

Work to do what? Are we talking about the cosmos's goal of observing itself, or maybe its goal of reducing free energy? Wouldn't not making this projection accelerate the heat death of the universe just as well as making it?

You might have said such a leap is permitted -- no constraint prevents it -- so it's inevitable that we do. But instead you said it's justified insofar as it works. And again I ask: what's the goal here that grounds this talk of what works? And whose goal is it?

So in all your examples, there are laid out certain constraints - which presumably are meant to achieve some effective action. And yet the actions look to defy them. A "wrong" procedure is employed to reach the apparent goal. — apokrisis

Nope. In each of my examples, the agent tasked with a specific goal chose instead to pursue a different goal, and in each case the initial goal could be taken as a means for achieving the new goal. But since there were other means available too, the agent can achieve the new goal without achieving the one they were tasked with.

The point was to show how the agent's judgment that what they did "works" could be faulty, unless some goal is taken as the goal relative to which a judgment of effectiveness is made.

Yet another example: suppose I'm trying to prepare for an exam, but my roommate decides to have some friends over for a party. When I remind him I have a test and complain about the noise, he says, "Well, you could go to the library, or -- you could come to the party. Check it out! I bought a copy of the test so you wouldn't have to study tonight."

See the difference here? The suggestion that I go to the library accepts my goal of studying effectively and suggests an alternative method of achieving that goal. The suggestion that I cheat substitutes the goal of getting a good grade, which admittedly was what I wanted to achieve by studying, but it bypasses what I have chosen to take as my immediate goal. If you want to judge what "works", you have to settle first what the goal is, and in many cases whose goal it is.

That is what I've been asking you to clarify. How your occasional appeals to a pragmatic "this, because it works" slot into your system. Works to do what? To study effectively or to get a good grade? And whose goals are we talking about?

the strangeness of your scenario — apokrisis

Here's another: A wedding planner tells one of the staff to put a certain flower arrangement on the dining table, and the ice sculpture on another table. The staff person decides to swap the indicated locations "because it looks better this way." Be that as it may, and whether the event planner agrees, the staff person cannot be said to have done what they were told.

But if the father's constraint was to find paid labour — apokrisis

It clearly was. Leave aside the robbery. Suppose the son just doesn't get a job. When his father asks, here is his answer: "If by 'get a job', you mean, did I follow the principle of least action, the principle of locality, and the second law of thermodynamics, all with an eye to Salthe's basic triadic system -- then 'yes'; if you mean, do I now have gainful employment, then 'no'." The son may effectively be accelerating the heat death of the universe, but he cannot be said to have a job.

You are relying on a highly artificial demarcation that seeks to stop us saying anything else about the situation. — apokrisis

"Did you get a job?" is a yes-or-no question. Humans do highly artificial stuff.

You've lost me. How could that have been the gist of your argument? — apokrisis

I posted it twice:

Whatever the merits of that view, and they may be considerable, it cannot be said that he succeeded at getting a job. — Srap Tasmaner

And thus the gist of my argument was:

did the son follow a procedure that is effective in achieving the goal of getting a job. — Srap Tasmaner

And the answer is "no", no matter what else we say about the situation.

***

My point was that effectiveness at achieving a goal at one level may not always count as effectiveness at achieving a goal on another level. It's a question about how exactly you attach the pragmatist appeal to the effectiveness of a procedure to the hierarchical/holistic/systems science analysis. I'm still not clear on that, which may be my fault.

Bonus example: I tell Sid to put his tools away; half an hour later I find little sister Hannah putting the tools away (no doubt because Sid threatened to decapitate her favorite doll unless she did). Sid's defence is "What does it matter who does it, so long as the tools get put away?" But it was his task to perform himself, not simply to make it that the task was performed. That kid's never going to learn how to be responsible ...

On your version - when forced to provide an intelligible rationale of the context in play - the son says it is all about the least action path to get that money. Jobs and robbery are not meaningfully distinct ... despite social norms that exist because of a larger scale social effectiveness. In the son's view, the father's attempt to draw a distinction is a quite arbitrary one on his own personal scale of being. Jobs or robbery is being claimed as a difference that should make no difference. — apokrisis

Then the "we" that should apply in your example becomes the social norms in play. A son that robs convenience stores is far more likely to come from a family and neighbourhood that robs convenience stores. The choice of a least action path to a goal would not really need much further justification.

But given your scenario, the father would be asserting some larger social norm as the "we" with the view on what is effective for that "we". We are law abiding and employed as that is a desire embedded at a cultural level, representing whatever happens to be functionally effective as a generalised habit. — apokrisis

Nope.

Whatever the merits of that view, and they may be considerable, it cannot be said that he succeeded at getting a job. — Srap Tasmaner

Full stop. I said nothing about norms, and gave no broader justification for the father's view. I thought of all that, because duh, but none of it is relevant to this single question: did the son follow a procedure that is effective in achieving the goal of getting a job. The answer to that is quite clearly, I submit, "no".

So I see no reason to give up the idea that effectiveness is tied to the achievement of specific goals, or carrying out specific tasks, and that even if those goals can be subsumed under other goals, effectiveness at achieving the higher goal does not pass through to the lower.

I heard a story once about an overnight shelving crew at a grocery store that knew the president of the company was visiting the next day: they weren't going to be able to finish getting the truck worked, so to avoid getting in trouble they threw a pallet of groceries in the big trash compactor.

The pattern here is similar: reinterpret the instruction to shelve everything as an instruction to leave nothing in the back. Reinterpret the instruction to get a job as an instruction to get money.

In some cases, that's fine. Maybe the instruction was unnecessarily specific. In the cases at hand, it's clearly a self-serving dodge. But none of this matters at all. Even in the acceptable cases, you'd be renegotiating the instruction; there's still no question at all about whether you did in fact achieve the goal you were directed to -- you didn't. Whether that's okay is another matter entirely.

The Cosmos only appears to be solidly there because it is - in some literal sense - observing itself. It exists as a globalised matrix of constraints on undirected local possibility.

Now the rejoinder is obvious. Quantum mechanics doesn't account for human feelings.

But I made the argument there too. Semiotics originated in phenomenology. It is rooted in the mechanics of human intelligibility. So it doesn't exactly leave the phenomenal out of it. Instead it accepts the full Kantian force of that and then builds back out so as to recover the noumenal - rescuing it via this idea of a core relational structure that acconts for intelligibility itself.

I can see the vulnerability that creates. Yes, we are projecting a view of ourselves as rational beings on to the apparent rational structure we see in the world. That is quite a leap of faith.

However - pragmatism again - suck it and see. Leaps of faith become justified to the extent they appear to work. — apokrisis

One question I'd like to raise regards this sense of effectiveness. Effectiveness is task-relative, to start with; there's no such thing as generic effectiveness. What's more, even if a given task slots into some hierarchy, effectiveness doesn't automatically cross those boundaries.

Example: father tells his son he needs to get a job; son goes out and robs a convenience store. When the father objects, the son's defence is that the whole point of getting a job was to get money, so he just got money a different way. Whatever the merits of that view, and they may be considerable, it cannot be said that he succeeded at getting a job.

So I'm wondering what task you have in mind when you reference this pragmatic sense of effectiveness, and whose task it is. Does this task belong to the "we" you reference, or to the cosmos?

Linguistic evolution appears to be at least partially Lamarckian, and that's interesting. I doubt that it is entirely so, but who knows.

Evolution connotes some kind of progression towards a goal — TheMadFool

Ick.

- adapting to the pressures of the language environment. — TheMadFool

I deliberately said nothing about adaptation.

Adaptation comes into biological evolution because there is an expectation that individuals better adapted to their environment are more likely to survive and (therefore) more likely to reproduce. That's not an exclusive deal, because sexual selection is a thing. But the point is always reproduction and surviving by being "fit" is just a means to an end. Natural selection predicts adaptation, for this very reason, but doesn't actually care. All that matters is that some individuals out-compete other individuals of their species in having their genes represented in the next generation. Reproduce more and you win. If adaptation helps, then that explains adaptation.

It's not clear to me that there is a similar story to tell about linguistic evolution. Reproduction in language, like talk -- because it is just talk -- is cheap. Almost preposterously cheap.

Of course there are selection pressures -- @frank gave a reasonable example of what that looks like, and that stuff is interesting. In nature, the penalty for a mutation that doesn't pan out is pretty harsh -- you die and you take all your genes with you, making it a little less likely that even the non-mutants will win in the next generation. In language, if a new usage (or whatever) doesn't catch on, oh well. It dies on the vine and the vine can keep churning out new grapes.

We're there? Sweet.

Btw, I live in the bible belt and a friend of mine who's a linguistics PhD has had encounters with people who accept linguistic but not biological evolution! He was in the position of arguing, "But you accept this with language, why ... ?"

It's a strange world.

It does seem more natural to say that my idiolect today is a descendant of, to start with, my idiolect, say, twenty years ago. At what point did I learn the word "idiolect", for instance?

What's clear though is that we're talking about differential reproduction, exactly as we do in biological evolution by natural selection, and that these changes in idiolects "sum" to changes in their species over time, and perhaps also to speciation. (The last being a can of worms I don't care to open right now.)

If it's the same process, then we would have to say that natural selection isn't too significant in language evolution (due to low population size). — frank

Except that it's obviously mainly natural selection. And you can see other stuff like drift when a population is geographically isolated, etc.

There are going to be some differences between biological evolution and linguistic evolution. Because the method of reproduction is so different, there isn't quite the same genotype/phenotype issue there is in biology, for instance. There's some stuff there I guess but that's linguistics I'm none too confident about.

To your specific point: I think you're matching things up wrong. A language is a species, not an individual organism. The counterpoint to an individual organism would be an idiolect, and maybe even a snapshot of an idiolect. In either case, you're not talking about small populations at all.

Some guy says to his friend, "That girl you were out with last night was 14-karat." Maybe he just saw a jewelry ad. His friend might like this expression and use it now & then. It could catch on in their circle and spread. Or maybe not. What's reproducing here is an idiolect that includes this usage of "14-karat". If nobody likes it, it could immediately go extinct. The idiolect is a member of the species English.

Is this making sense? It was my understanding that this isn't remotely controversial within linguistics, but I haven't done any googling since this thread started.

There's a structural isomorphism between the two processes, biological and linguistic evolution, so it's more than an analogy.

Look at how Latin (or camp Latin) speciates into the romance languages. You're looking at an accumulation of changes in frequency of pronunciations, meanings, spellings, etc.

Where biology relies on biological reproduction, languages have something else, this is true. But usages and pronunciations and the rest do get reproduced or not within a community. Some words catch on. Some meanings, etc. Some don't. And over time the language changes. Languages have descendants, go extinct, all the rest.

There are multiple phenomena you can look at: changes in pronunciation, in meaning, in whatever the word is for degree of "elevatedness", etc. All of these follow a similar pattern. Changes in the frequency with which a community is using an allophone, for example, slots into the math where biology would have changes in the frequency of an allele. And that's why languages form family trees much like species. A hypothetical language like proto-indo-european is a most recent common ancestor just like biologists talk about.