Episode Transcript
Transcripts are displayed as originally observed. Some content, including advertisements may have changed.
Use Ctrl + F to search
0:00
Hey, Big Biology listeners! A reminder that we're
0:02
moving into a more intensive period of fundraising
0:04
over the next few months, so you'll notice
0:06
a couple of changes to our show. First,
0:08
if you listen to us on Spotify, you'll
0:11
now hear two or three short ad breaks
0:13
during the hour. Remember, we're a non-profit entity,
0:15
and we don't want to insert commercials into
0:18
the show, but to keep it going, we
0:20
have to use all options. Second,
0:22
if you're a regular listener of Big
0:24
Biology, consider donating through our website, or
0:26
even better, sign up to be a
0:28
patron by making a monthly donation
0:31
through Patreon. If enough listeners donate the
0:33
equivalent of what an espresso or two
0:35
costs per month, you can help us
0:37
have stable funding that will keep the
0:39
show going well into the future. And
0:41
remember, Art, Cam, and I donate our
0:44
time. But producer Molly, interns Dana and
0:46
Keating, and our webmaster Steve, they
0:48
all depend on your support. Great.
0:51
So either go to
0:53
our website, www.bigbiology.org, and
0:55
make a donation, or
0:57
go to patreon.com/bigbio and become a patron
0:59
for just a few dollars a month.
1:02
Patrons also get cool insider stuff, like
1:05
access to behind-the-scenes audio from our guests
1:07
about their lives, their hobbies, and their
1:09
careers. Now, on to the show. Marty
1:22
and Art, it's good to see you
1:25
two philosophers today. Philosophers? Norwegian
1:27
Winter got you confused, Cam.
1:29
We're biologists. Wait, but PhD
1:32
stands for Doctor of Philosophy.
1:35
Technically, Cam's right. We are philosophers
1:37
in the sense that all sciences
1:39
are rooted historically in philosophy, and
1:42
that meant the pursuit of wisdom,
1:44
scholarship, and expertise. Fail. But
1:47
there is a discipline where the
1:49
two meet. The philosophy of biology
1:51
explores fundamental topics like how the
1:53
meaning of terms and critical assessments
1:55
of assumptions have changed through time.
1:57
And if you're a regular big biology listener, you
2:00
probably heard past episodes with philosophers
2:02
of biology like Dan Nicholson, Dennis
2:04
Walsh, and many others. I
2:06
think most practicing biologists, myself included, rarely
2:08
read the philosophy of biology literature, but
2:11
those are some of my favorite episodes
2:13
because they've made me stop and reflect
2:15
on ideas and concepts that I often
2:17
use but never really thought deeply about
2:20
their meanings. In a 2019
2:22
article in the Proceedings of the National
2:24
Academy of Sciences, Lucy Laplan and colleagues
2:26
argued that philosophy plays a critical role
2:28
in the sciences. An example they highlight
2:30
is a critique of the cell versus
2:33
non-self idea in immunology. Philosophers pointed out
2:35
that the idea of self breaks down
2:37
will be recognized that all organisms are
2:39
made up of communities of non-self symbionts
2:42
and the shift has real-life implications
2:44
for developing medical treatments. Thank you
2:46
philosophers. One active area
2:48
in the philosophy of biology is
2:50
evolution by natural selection. Concepts
2:53
like fitness, selection, adaptation, contingency,
2:55
and determinism pop up all
2:58
the time. But talk
3:00
to philosophers of biology and you find out that
3:02
their meanings are far less clear than you would
3:04
think. For example, think about the
3:06
phrase survival of the fittest. When
3:09
we define fitness as the ability to
3:11
survive and reproduce, the phrase becomes tautological
3:13
because we're saying essentially survival of those
3:16
who survive. This is why the philosopher
3:18
Eliot Sober has said we need a
3:20
better definition of fitness. Another topic we've
3:22
covered a lot recently on the show
3:24
and that also shows up in a
3:27
lot of philosophical analyses is
3:29
agency. Agency has been
3:31
defined very differently inside and outside of
3:33
biology. Agency also carries a lot of
3:36
baggage which puts a lot of biologists
3:38
on edge. When you say
3:40
that organisms have purposes and goals, it
3:43
opens the door to conscious intention
3:45
and even supernatural effects and forces.
3:47
Yet even if we recognize that
3:49
most of what organisms do to
3:51
survive and reproduce is done unconsciously,
3:53
we often use casual and
3:56
sometimes clumsy language. Think about the cuckoo,
3:58
a bird that lays eggs in
4:00
the nest of other species. The cuckoo egg
4:02
typically hatches first, and the cuckoo chick then
4:04
pushes the host eggs out of the nest.
4:07
We might casually say that the cuckoo chick
4:09
does this because it consciously wants all the
4:11
food for itself. We often use
4:13
this kind of language as shorthand to help
4:15
us explain why the chick behaves the way
4:17
it does, even though it
4:19
may not be making a conscious choice.
4:22
A more traditional explanation would say that
4:24
this behavior evolved by natural selection acting
4:26
on heritable variation in the context of
4:28
an evolutionary arms race with its host.
4:31
So, while there is a growing interest
4:33
in the notion of agency in evolutionary
4:35
biology, there's still a lot of confusion
4:37
and debate, even among us, about what
4:39
it means, as well as its relationship
4:41
to fitness, selection, and adaptation.
4:44
Our guest today is Samir Akasha, who's
4:46
a professor of philosophy of science at
4:48
the University of Bristol in the UK.
4:50
His 2006 book, Evolution and the Levels
4:52
of Selection, won the Lakatos Award. Today,
4:54
we talk with Samir about his 2018
4:56
book entitled Agents and
4:59
Goals in Evolution. In contrast to
5:02
our past discussions of agency, Samir takes
5:04
a different approach. He
5:06
asks in the book, quote, Does
5:08
thinking about organisms as agents have
5:10
a genuine scientific role to play in
5:12
our understanding of adaptive evolution, or is
5:14
it simply another application of anthropomorphic
5:16
language? In our chat, we
5:18
look critically at the different definitions of
5:21
agency in biology. How is it different
5:23
from phenotypic plasticity? And do
5:25
we separate agency and fitness? And
5:27
what about agency below the organismal level?
5:29
Could a gene be considered an agent
5:31
if it acts selfishly to increase copies
5:33
of itself, like in cases of intergenomic
5:35
conflict? And what about groups
5:38
of humans or even groups of different
5:40
species that exhibit cooperative, goal-directed behavior? Do
5:43
groups have agency distinct from individual
5:45
members? Samir argues that
5:47
although we gain insight by
5:49
anthropomorphizing some adaptive traits, there's
5:51
still value in understanding how agency
5:53
influences evolution, even if we don't
5:55
know about consciousness in that organism.
5:58
We also discuss whether the process of
6:00
evolution by natural selection represents a form
6:02
of agency. This might sound odd,
6:05
but recall that Darwin himself depicted natural
6:07
selection with a language of intent, as
6:09
a background process that's always scrutinizing
6:12
populations, keeping the good and eliminating
6:14
the bad. We debate
6:16
some really difficult problems like whether agency
6:18
can evolve, is there heritable variation for
6:20
it, and how would we measure that?
6:22
Something we've been thinking about a lot,
6:25
but I'm not sure we've totally resolved.
6:27
So stay tuned for a philosophical assessment
6:29
of agents, goals, and evolution. I'm Cameron
6:31
Gallenbore. I'm Art Woods. And I'm Marty
6:33
Martin, and you're listening to BIG Biology.
6:51
Samira Kasha, thanks so much for joining us on
6:53
BIG Biology today. It's my pleasure.
6:55
We're super excited to talk to you
6:57
today about your recent book, Agents and
7:00
Goals and Evolution. Maybe to
7:02
begin, could you start off by
7:04
defining what is a
7:06
biological agent? Well, that's a
7:08
complicated and difficult matter. Let
7:11
me give you my take on it. As
7:13
I see it, the place we need to start is
7:15
by asking pretty much the question you did. I mean,
7:18
what do we mean by the term agent? And
7:20
why do people think that this is a good concept
7:22
or term to use when thinking about
7:25
biology? And I think that's an important
7:27
question to ask, because there's this flurry of
7:29
interest in the notion of
7:32
agency in biology, as many of
7:34
your past contributors have discussed, and as you'll
7:36
be aware, and as a lot of philosophers
7:38
have joined this discussion too. But
7:40
on the face of it, I mean, that's a
7:43
bit odd, because if you open any biology textbook
7:45
in any branch of biology, I
7:47
very much doubt that the term agent is going
7:49
to appear in the index. So
7:52
that basically raises the question, well, why should
7:54
this notion be so important for sort of
7:56
philosophical reflections about biology if the science itself
7:58
is a good thing? doesn't seem to have
8:00
any use for it. So that's I think the first sort
8:03
of skeptical question I would ask and next
8:05
move I would make is then to say well if
8:07
people do want to talk about this let's do the
8:10
sort of standard thing we do in philosophy of saying
8:12
well what exactly do you mean? It's
8:14
always a good question to ask. And
8:16
as I see it there are I mean there are
8:19
a number of different motions of agents at work
8:21
both in different sciences and
8:23
in vernacular English. So
8:25
in my book I distinguish between
8:27
four different notions of
8:29
age and as a sort of prelude to
8:32
the discussion. The first is
8:34
what in in traditional philosophy we
8:36
would call an intentional agent. Someone
8:38
or thing that engages in reasoning
8:40
and in thinking and in deliberating
8:42
about what to do and choose
8:45
these courses of actions because of particular
8:47
goals that they want to achieve. Where
8:50
the goal is something that's mentally represented
8:52
by the agent. So that means
8:54
that at the very least you've got to have a mind
8:57
of some sort. Now what exactly that
8:59
I mean that's clearly intended to include
9:01
humans. Whether it includes what
9:03
else it includes you know whether
9:05
it includes all vertebrates or even
9:07
some invertebrates or even possibly microbes
9:10
it could be debated. But that as I see
9:12
it is the sort of traditional philosophical notion of
9:15
agent. An agent engages not
9:17
just in behavior but in action where
9:20
action means that they're deliberately trying to
9:22
achieve the goal. And that's what
9:25
I think of as the most
9:27
restrictive notion of agent. The literal
9:29
applicability of which is the least
9:31
if you see what I mean.
9:33
But then in different sciences we
9:35
find broader notions of agents. So
9:37
for example in AI one finds
9:39
frequent reference to talk about intelligent
9:41
agents. Where that basically means
9:43
anything you could be a control system
9:45
or piece of software or a thermostat
9:47
that doesn't always do the same thing.
9:50
That does different things depending on
9:52
the circumstance that it's in. So
9:54
typically we'll sense the environment somehow
9:57
and change its behavior depending on what
9:59
it says. I mean, so implement
10:01
stimulus response conditionals. It's an
10:03
extremely inclusive notion of agents,
10:06
that of course. Similarly,
10:08
but differently, one finds in the
10:10
economics literature talk about rational agents
10:12
and a lot of talk about
10:14
the rational agent model, where
10:17
essentially that means an
10:19
agent who engages in consistent
10:21
choice, so behaves as if
10:24
they're trying to maximize some utility
10:26
function or something like that, as
10:29
whose choice behavior is
10:31
consistent, satisfies conditions such
10:34
as transitivity, things like that. Again,
10:36
that's a broad notion of agents
10:38
in that there are no psychological
10:40
requirements. I mean, any organism
10:42
that engages in choice or behavior
10:45
of any sort could satisfy
10:47
or fail to satisfy those conditions on
10:49
being a rational agent as
10:52
understood in rational choice theory
10:54
and economics. And then
10:56
the final notion of agent that I operate with,
10:59
or that some people operate with that I distinguish,
11:01
is what I have called minimal agent.
11:04
And that just means something that
11:06
does something. So an agent
11:09
in this sense is just something that does
11:11
something rather than something to
11:13
which things happen, if you like. And
11:16
how exactly you cash that out, I
11:19
think is a difficult philosophical question. But
11:22
I mean, examples can make it clear. So to
11:24
borrow one from the philosopher Fred Dretsky that I
11:26
like to use, he said, look,
11:28
contrast these two things. A rat
11:30
moves its paw in
11:33
order to press a lever to get a pellet
11:35
of food. And an
11:38
experimental scientist puts the hand into the
11:40
cage and forcibly moves the rat's paw.
11:43
In the first case, the rat moved its
11:45
paw. In the second case, its paw was
11:47
moved. And
11:49
that he says illustrates the difference between
11:52
doing something and having something happen to
11:54
one. Sort of active versus passive. Yeah,
11:57
exactly. And then
11:59
obviously... the next thing to say about that is, well,
12:01
what, how do we cash that out precisely? What does that
12:03
really mean? And some philosophers have
12:05
thought that the answer simply lies in the
12:08
question of whether the cause was internal to
12:10
the agent or not. You
12:13
know, so if a bird flies, then
12:15
that's one thing, a hurricane, force, wind
12:18
blows the bird. That's another thing. In
12:20
one case, the cause of the, the
12:22
motor output is internal. In the second
12:24
case, there's an external cause. So
12:27
that, that's what I call the most minimal notion of
12:29
agents entirely. So those
12:31
are the four notions I operate
12:33
within some sort of rough hierarchy,
12:35
the most demanding, the intentional agent,
12:37
the least demanding, the minimal agent,
12:39
and then the intelligent agent and
12:42
the rational agent somewhere in between
12:44
those two extremes. And
12:46
then the next question to ask is,
12:48
well, which event, if any of those
12:50
notions applies to biological systems, to
12:53
which biological systems do they apply and
12:55
how do we know and why? But here
12:58
a complication comes in because I maintain, and
13:00
this is my, I guess my real interest
13:02
in my book, the much
13:04
talk of agency in the
13:06
biological realm is metaphorical rather
13:08
than literal. Particularly
13:10
in evolutionary thinking, I
13:13
claim that there's a lot of appeal
13:16
to the intentional notion of agent, the
13:18
most demanding of all, that
13:20
I think of as in many ways, the core notion,
13:23
where the agent must have a mental representation of
13:25
the goal. But the concept
13:27
is only applied metaphorically. When we say
13:29
things like the bacterium swims up an
13:31
oxygen gradient, because it wants to get to
13:33
an area of higher oxygen concentration or something like that. Yeah.
13:37
So let's, let's unpack that first one a little
13:39
bit, because I had a question lined up about
13:41
that very thing. So, you know, in the, in
13:43
the book, you make it clear that for this
13:45
more restrictive idea of agency, that it requires a
13:49
mind and a mental representation.
13:51
And at least like among
13:53
those of us, the hosts that are talking
13:56
about agency, Marty and I have sort of kicked this
13:58
back and forth. I think our concern of
14:00
agency is a little bit broader
14:02
and it encompasses basically any complex
14:04
system that's able to receive
14:07
information, process it in some way,
14:09
and make a decision. And that
14:12
doesn't invoke anything about nervous
14:14
systems or minds or mental representations. You
14:16
certainly could have biochemical networks that are
14:18
making decisions. You could make, you know,
14:21
have sort of genetic regulatory networks that
14:23
are making decisions. And so I think
14:25
that that immediately expands this
14:27
idea of agency down to very
14:30
simple organisms like an, you know, an E.
14:32
coli deciding whether or not to make the
14:34
machinery to metabolize lactose or
14:36
not. So in your conception, why do you
14:38
have to have a brain and a mental
14:40
map? Well, I don't say that
14:43
you do. I only say that
14:45
you do in order to be a
14:47
literal agent in one sense of the
14:49
term agent. But I entirely hear your
14:51
line of argument. And I think it's
14:53
one I'm sympathetic to that says, look,
14:56
there are, you know, other, so what
14:58
you might call hallmarks of agency are
15:00
found in pretty much all complex
15:02
systems, biological and maybe
15:04
even non-biological, certainly including
15:07
cells and possibly subcellular
15:09
constituents and molecular
15:11
pathways and things too. However,
15:14
I would say that that involves a somewhat
15:16
different notion of agents. So
15:18
I would actually say that's closer to the
15:21
intelligent agent notion of AI. So
15:23
basically, what we mean is that
15:25
we've got something that receives
15:28
information from the environment
15:30
and modifies its behavior
15:33
as a consequence. So in short, it doesn't always
15:35
do the same thing. What it does depends on,
15:37
on circumstance, if
15:40
you like. But then I mean, then
15:42
if that's right, then it raises the
15:44
question, do we just mean behavioral plasticity?
15:47
When we talk about agency in that sense? Exactly
15:50
what I wanted to ask. Is
15:53
this a continuous grade from intelligent
15:55
to intentional? I mean,
15:57
and is there some way to find a clear
16:00
demarcation between these kinds of
16:02
things? Well, presumably, I mean,
16:04
the intentional one is
16:07
a relatively late arrive on the
16:09
evolutionary scene. In that,
16:11
you know, however widely we think the
16:14
true intentional agency is found in
16:17
the animal world, and, you know, cognitive pathologists
16:20
have, you know, often debated, for
16:22
example, whether corvids are capable of genuine
16:25
intentional behavior or not, and opinions divide on
16:27
that issue. But what's very, very clear is
16:30
that, well, I would
16:32
say at least the vast majority of
16:34
organisms, including all invertebrates, are
16:36
not literally intentional agents. I mean, if we,
16:38
we might describe them as if they were,
16:40
when we say, you know, if the cell
16:42
wants to achieve such and such an end,
16:44
but that's, to deliberately
16:46
anthropomorphize, and maybe for good reason,
16:49
but it's not literally true. It's a
16:51
bit like saying the suitcase doesn't want to close, you
16:53
know, it might be useful thing to say, it's not
16:55
literally true. Right. What do you think
16:58
Lars Chitka might say about that? I'm not sure
17:00
if you're familiar with his work, but we spoke
17:02
to him quite a while ago. So
17:04
he's very interested in insect cognition. And
17:06
he's done really amazing work on all
17:08
sorts of bees in terms of their
17:11
capacity for solving puzzles. One
17:13
of the really neat stories that he tells is,
17:15
you know, how do bees decide,
17:17
decide to
17:19
build the nest that they do, and
17:22
the configurations that they can take, you know,
17:24
it's not like it's a sort of hardwired
17:26
always the same kind of
17:28
nest, it's very responsive to
17:31
the conditions of the environment. And he's
17:33
done incredibly sophisticated experiments to show the
17:36
amazingly elaborate kinds of outcomes from, you
17:38
know, presumably a simple cognitive
17:40
system. Yeah, no,
17:42
I mean, I think it's a genuine challenge
17:44
that I mean, at some level, it's got
17:46
to be an empirical question, I think. One
17:49
way to look at it would be to say,
17:51
well, what's to be gained, this is a formulation
17:53
inspired by Daniel Dennett, where he says something
17:56
like, heuristically, what would we gain, if
17:58
anything, by treating some
18:01
organism or collection of organisms in
18:04
the B case perhaps as an
18:06
intentional agent as if it really was trying
18:08
to achieve some goal and had
18:11
beliefs about the world and processing
18:13
information and so on. I
18:15
mean is that a useful perspective or not?
18:17
I mean the answer might very well be
18:19
yes in that sort of example but I
18:21
take it that there are examples where it
18:24
would not where we have a
18:26
purely mechanistic understanding of
18:28
the movement of some microbe
18:31
or something like that or the
18:33
growth pathway of some plant for
18:36
example. I mean it would be seem
18:38
superfluous to give a psychological explanation
18:41
of plant growth towards the
18:44
light or something then. I
18:46
mean we have a fairly
18:48
good mechanistic understanding of why the
18:51
plant responds to light
18:53
that way and if someone wanted to say well
18:56
no it's because it wants to get more sunlight
18:58
or something or wants
19:00
to grow taller than its neighbor. That would not be,
19:03
I take it scientifically useful. So one test
19:05
is just the scientific utility. Now
19:08
some philosophers and others would oppose that
19:10
methodologically by saying look the question of
19:12
whether something is useful is distinct from
19:14
the question of whether it's true and
19:16
that may be but nonetheless I
19:18
think if it isn't useful at all then
19:20
that's some indication that it's not literally true.
19:23
So as I see it I mean I wouldn't at
19:25
all insist that we know that it could only be
19:28
in vertebrates that we find
19:30
the hallmarks of true intentional agency. I
19:32
mean another point though to bear in
19:34
mind is that I mean
19:36
I started when I defined intentional agency by saying
19:39
that there really is a mental representation of the
19:41
goal. The organism really
19:43
it's not just as if it was trying
19:45
to achieve a goal it really is. But
19:48
you see if you talk instead about decision then
19:50
that's perhaps a less psychologically
19:52
demanding notion than having a
19:55
mental representation of the goal in that I
19:57
think it's perfectly sensible and plausible to say
20:00
you know, even of a developing organism that
20:02
it needs to decide which pathway to
20:04
go down. So developmental
20:07
regulatory networks can make decisions, chemical
20:10
networks can make decisions. Yeah, I mean,
20:12
you know, some people might insist that
20:15
the literal meaning of the word decision is a
20:17
conscious decision made by human. It's sort
20:20
of inherently mental. Yeah, right. Yeah, but
20:22
I don't see that. I don't see
20:24
that that's either the standard
20:26
way that the word is used in science these
20:28
days, or has anything
20:30
particular to recommend it sort
20:32
of theoretically. So I'm quite happy for
20:34
decisions to be a pretty broad thing. I'm
20:37
very happy with the idea then that, you
20:40
know, all organisms, including microbes
20:42
make decisions and maybe decisions
20:44
are made at the level
20:46
of regulatory networks and, you
20:48
know, subcellular levels too. But
20:51
if so, then I would say that
20:53
decision making and having a mental representation
20:55
of the goal come apart.
20:59
Even that the former is taxonomically widespread, the
21:01
latter isn't. So if
21:03
we can circle back a little bit
21:05
to the concept of plasticity and flexibility,
21:08
for I'd
21:10
say most practicing evolutionary
21:12
biologists, the capacity
21:15
for being
21:17
flexible and whether it's
21:19
at the level of development in response
21:22
to some environmental cues or in the
21:24
context of behavior changing, say in the
21:27
presence and absence of a predator, a
21:30
lot of those types of responses
21:32
are captured currently under the sort
21:34
of framework of phenotypic plasticity. And
21:36
so just based on your
21:39
description, it seems that being
21:41
plastic and flexible
21:44
is inherent to being
21:47
agential in your response. Yeah, I would
21:49
agree. In all of the senses of
21:52
agency, in all of the four senses
21:54
that I've operated, I've distinguished. So are
21:56
they so intertwined that they are essentially
21:59
capturing? the same phenomenon or can we
22:01
sort of decouple them from one another?
22:04
Yeah, I know. I mean, it's tricky. I
22:06
mean, I do think that, you
22:09
know, the notion of phenotypic plasticity and
22:11
behavioral plasticity, that being a subset of
22:14
phenotypic plasticity, you know, given that not
22:16
all phenotypes of behaviors, you
22:19
know, captures a lot of what people
22:21
who talk about agency are
22:23
trying to say, and probably in a
22:26
more sort of acceptable way
22:28
to many practicing scientists. And I think
22:30
that's exactly the source of some of
22:32
the opposition to the language of
22:35
agency that one finds these days
22:37
where people say, look, you're just using
22:39
this mystical-sounding term, which the
22:41
cash value of which is something that we could say
22:43
anyway. Right. Now, you see,
22:45
I think that's a valid criticism of some
22:47
projects that involve the notion of agency, but
22:50
not of mine. In
22:53
some theorists, including some of
22:55
my colleagues in philosophy of
22:57
biology, seem to have convinced
22:59
themselves that there's a deep
23:01
and interesting sense in which
23:04
some organisms or maybe all organisms
23:06
and maybe other biological entities
23:09
too, are agents. And
23:11
their task is to figure out what that sense is,
23:14
subject to the constraint that it
23:16
better be interesting, and better not
23:18
just reduce the behavioral plasticity. But
23:21
that seems to me an odd way to proceed. So
23:24
that's why I'm one of the people who's
23:26
somewhat skeptical about the notion, despite having worked
23:28
on it myself. But my project
23:30
is a somewhat different one in that
23:33
I was, in writing my books,
23:35
specifically interested in a mode
23:37
of thinking and speaking
23:40
that is peculiar to evolutionary
23:42
biology, I think, which
23:45
involves deliberately
23:47
anthropomorphizing for
23:49
the purposes of achieving evolutionary insights.
23:52
And one example of which is
23:54
that the practice of taking
23:57
the adaptive function
24:00
or rationale of some phenotype,
24:03
but particularly a behavior, but not
24:05
necessarily, in the sense that the
24:07
reason why it's evolved. And then
24:09
treating that as if it was the
24:11
organism's own goal, a
24:14
literal goal mentally represented, even when
24:16
it really isn't. And
24:18
I give many examples in my book
24:20
of this practice actually at work in
24:22
evolutionary biology, manifest by the use of
24:24
intentional language. So when we say things
24:26
like that the female rat kills its
24:28
injured young, because it knows that they
24:30
won't survive, it doesn't want to waste
24:32
resources on them, for example, there
24:35
we explicitly employ, you know,
24:37
vocabulary that has its natural home in
24:40
relation to conscious deliberate human behavior, by
24:43
way of giving an evolutionary explanation, or an
24:46
adaptive explanation of the infanticidal
24:48
behavior of the female rat.
24:51
And so that was the phenomenon I was
24:53
specifically interested in the use of intentional vocabulary
24:57
in an evolutionary context. And
25:00
so I took my project, I mean, I
25:02
approached it very much in the spirit of
25:04
a kind of anthropologist of science, you know,
25:07
my starting point was, look, there is this
25:09
pre existing practice out there in the scientific
25:11
literature, that's really unusual and
25:13
philosophically puzzling. And what
25:16
on earth could be the point of talking
25:18
that way, given that it seems to willfully
25:20
invite a confusion of proximate and ultimate? And
25:23
given that it seems to just be a
25:25
form of anthropomorphism, could it really be
25:27
doing any scientific work? That was
25:30
the question I started from. And
25:32
I think of that as a somewhat
25:34
different project, indeed, an orthogonal project, to
25:36
the one that many people, you know,
25:39
both biology and philosophy, who talk about
25:41
agency are ultimately engaged in. So
25:43
I want to push back a little bit, because you
25:46
made a great case, and this adaptationist
25:49
program has been and is pervasive in
25:52
biology. But there's this particular word that
25:54
I hear you using that I'd like
25:56
you to address, and it's mental, right?
25:59
So An important
26:01
qualifier of agency
26:03
that you write about is integration.
26:06
So it's not just the
26:09
potential to perceive the problem or to sort
26:11
of think about the problem, but then the
26:13
system does something about it. So can
26:16
you expand the definition of mental? And
26:19
more importantly, it seems to me to
26:21
be leaning on mental as
26:23
the qualification of agential and justification
26:25
of how to think about that.
26:27
That seems akin to the anthropomorphizing
26:30
and adaptationist perspective. It's a
26:32
move that allows you to
26:34
do other things, but I
26:37
don't know. I guess I'm just not on board that
26:39
mental is something special
26:42
because you have to have integration. And as Art
26:44
was saying a minute ago, there's a lot of
26:46
different systems and even elements of systems that integrate.
26:49
Yeah, no, I mean, I agree that
26:51
integration and mentality are not
26:53
exactly the same thing, certainly.
26:55
So I mean, what I'm getting
26:57
at when I use the term mental is
26:59
that just the assumption that
27:02
psychological descriptors like believes,
27:06
wants, tries, apply
27:09
literally in the human case and
27:12
quite possibly in other non-human
27:15
cases too. But
27:17
don't apply literally in all cases
27:19
where there are sort of hallmarks
27:21
of agency in the sense of
27:24
flexible behavior, plasticity,
27:26
decision making, information
27:28
integration, all of that. Now,
27:31
I mean, that move could be
27:34
contested, I agree. That's not self-evidently
27:37
true. But I think
27:39
that's the natural default assumption that this is
27:41
the home of this vocabulary. It's
27:44
really weird to my mind to think
27:46
that the psychological language of
27:49
that sort literally characterizes what an
27:51
earthworm does. I
27:53
just don't think it's true. And
27:55
that's fair, but I think that's where I'm
27:57
coming from. I mean, to just sort of…
28:00
distinguish these forms of agency using
28:02
human cognition as
28:05
sort of a frame for how we think about all
28:07
of these other things. I sort of approach it from
28:09
the other direction. Life as a
28:11
complex system, one of its manifestations ends up
28:13
being the human brain, but
28:15
even the vertebrate immune system can learn
28:17
in an exquisite way that
28:19
has nothing to do with how nerves
28:21
learn. I mean if we had a
28:23
better grasp on what consciousness was and
28:25
the interplay between consciousness and mental,
28:29
maybe I'd feel a little bit more comfortable, but because those
28:31
are such abstractions too, I don't know, it
28:33
feels like a strange place to start. Right,
28:36
I see. Yeah, I mean why
28:38
do I start in that place? You see, I
28:40
mean I entirely agree with you that, for example,
28:42
the adaptive immune system learns. Yeah,
28:45
that seems that seems absolutely clear. But
28:47
why do I start with the mental?
28:50
I mean I think it's because my
28:52
specific interest is in this anthropomorphizing tendency,
28:55
or at least in that book,
28:57
that's the thing I'm specifically interested
28:59
in, is this anthropomorphizing tendency within
29:01
evolutionary biology. And my overarching question
29:03
in that book was, look, is this tendency
29:06
to use the language of intentional
29:08
psychology in an evolutionary context
29:10
in order to cause adaptive
29:13
explanations, is there any
29:15
real good reason to do that? Or
29:18
is it just a sort of casual
29:20
anthropomorphic bias? Does it pick out some
29:23
genuine class of phenomena? Or is it,
29:25
does it really just more reflect the
29:28
fact that the investigators are humans
29:30
themselves and haven't entirely always managed
29:32
to shed themselves of their anthropomorphic
29:35
bias? But in terms
29:37
of thinking about the origins of agency
29:39
itself and the origins of intentionality
29:42
and consciousness, I mean, I
29:44
think I agree with you
29:46
that precursor forms of all
29:48
of these things are found
29:50
in a much simpler biological
29:52
systems and that somehow
29:54
or other, you know, the
29:56
evolutionary process gave rise to creatures
29:58
whose cognition was sophisticated. sophisticated enough,
30:01
for whatever reasons exactly, that
30:04
they could actually mentally represent
30:06
their goals. Yeah. So Samira,
30:08
I gathered from your book that you
30:11
sort of slot these different ideas about
30:14
sort of levels of agency into what you
30:16
call type one and type two. And
30:18
type one is about organismal agency
30:20
itself, which I think, you
30:22
know, now many biologists are starting to hear
30:24
about and having strong reactions to.
30:27
Type two is this idea
30:29
that the process of evolution
30:32
itself and natural selection can
30:34
be agential and or
30:36
that it may appear to be
30:38
agential because of the sort of
30:40
psychological and intentional language that's used.
30:43
And I would say like it feels like
30:46
your book is largely, you know, in
30:48
the end dismisses this idea that type
30:50
two agency is important and actually captures
30:52
something real that's happening in biology and
30:54
that that we're sort of, in
30:57
a sense, taken for a ride by
30:59
the language that we use to describe
31:02
type two agency. Yeah, no, that's
31:04
exactly right. So it was, yeah, my sort of
31:06
simple one line argument is that
31:08
type one good in certain cases,
31:11
scientifically justifiable, despite being
31:13
anthropomorphic, type two,
31:16
generally not good, anthropomorphic,
31:18
but bad. So type two in
31:20
at least the paradigm sort
31:22
I have in mind involves, you
31:24
know, thinking of the evolutionary process,
31:26
as you say, or the process
31:28
of natural selection as a form
31:31
of agency, if you like. And Darwin
31:33
himself was the first to do this
31:35
in a number of passages in the
31:37
origin of species, where he says natural
31:39
things like natural selection is daily and
31:42
out, he's scrutinizing the world, looking
31:44
for any advantage and preserving
31:47
what's good, eliminating what's bad. So
31:49
that is quite deliberately using this
31:52
intentional language to characterize the
31:54
process of natural selection itself.
31:57
Now, Darwin went on to say, look, these are
31:59
just metaphors, but don't worry about it. It's
32:02
not a big deal. And I think
32:04
in fact, he was cavalier on that point. Because,
32:06
I mean, one of the
32:08
things that people didn't realize in the didn't
32:10
understand in the earliest days
32:12
of post Darwin was that the
32:15
process is, of course, mechanistic and
32:17
doesn't have any foresight, there is
32:19
no goal sorts, which evolution is
32:22
striving or anything like that. And, you
32:25
know, any attempt to think otherwise
32:27
is really is to misunderstand Darwin's
32:30
achievement. But nonetheless, I do
32:32
think that sort of vestiges
32:34
of that type two form
32:36
of agential reasoning and thinking
32:39
are still alive today, in
32:41
fact, because you see, I think
32:44
that we very often treat the
32:46
simplest sort of natural selection as
32:48
if it was paradigmatic. And
32:51
so by I mean, the simplest sort of natural
32:53
selection is just one we've got a number of
32:55
variants in a population in a fixed environment, with
32:58
no frequency dependence. And
33:01
the evolutionary process when driven by
33:03
natural selection simply involves fixation of
33:05
the best variant and elimination of
33:08
the worst ones and the, you
33:10
know, monotonic increase of mean
33:13
population fitness until it reaches
33:15
a maximum so pushes the
33:17
population up the adaptive landscape, all
33:20
ideas that are, you know, have their place.
33:22
And are useful, but rest on, you know,
33:24
that's only the very simplest case. And,
33:26
you know, it's all the evolutionists know,
33:29
those things hold true under restrictive model
33:31
assumptions that almost never apply. And in
33:34
the real world. And so I think
33:36
the tendency to adopt that
33:38
language leads us to
33:40
think that the simplest form of
33:42
natural selection as a sort of
33:45
simple optimization process is
33:47
the paradigm example. And
33:49
obviously people who really know the science
33:51
know that that isn't true. But
33:54
nonetheless, I think a vast number of
33:56
people who sort of know a bit about evolution and
33:58
natural selection, don't understand. understand that that
34:00
isn't true. And the use
34:02
of this intentional language, thinking of
34:05
the evolutionary processes, you
34:07
know, either driven to achieve some
34:09
goal or in any case as directed
34:12
in some sort of way or as pushing
34:15
populations up landscapes by the steepest
34:17
path, that sort of thing. Those
34:20
mistaken notions go hand in hand, I
34:22
think, with the use of these
34:25
anthropomorphic intentional descriptors to characterize the
34:27
process of evolution by natural selection.
34:29
So that's why I say type
34:31
too bad. But I want
34:33
to say type one, which involves
34:35
thinking of an actual evolved entity,
34:38
typically an organism, but possibly a
34:40
group or maybe even a gene
34:43
as akin to an agent, and
34:45
as trying to achieve a goal can
34:47
be in certain cases, a
34:50
useful mode of description.
34:52
I try and make this argument
34:54
that although anthropomorphic, it does genuine
34:56
work in that it picks out
34:58
a real phenomenon that, you
35:00
know, certainly could be described otherwise, but that
35:03
there is a real reason to describe this
35:05
way. And that's when I start
35:07
talking about integration and
35:09
absence of internal conflict as
35:12
being a prerequisite for thinking of
35:14
an organism as agent
35:16
like. Hey
35:24
there! Did you know Kroger always gives
35:26
you savings And
35:30
when you download the Kroger Ap, you'll enjoy
35:32
over five hundred dollars in savings every week
35:34
with digital coupons. And don't forget few points
35:36
to help you save up to one dollar
35:38
per gallon at the pump. Wanna. Save
35:40
even more with a boost membership you'll
35:42
get double few points and free delivery.
36:01
Yeah, so this is a good place to
36:03
really just ask the question, are genes
36:06
agents? Especially in
36:08
the context of thinking about transposable
36:11
elements and gene
36:13
for altruism or other
36:16
kinds of inter-genomic conflict, it
36:18
seems like our language
36:21
that we would use and our
36:23
theory to describe why there would
36:25
be selection for or against the
36:27
spread of a gene within a
36:29
genome fits within this concept
36:31
of an agent. Yeah, no, I
36:33
agree with that. And I think, I mean,
36:35
the broad story I'm trying to tell about
36:37
agency and the circumstances in which it's a
36:40
valid concept or useful concept
36:42
at least to employ in an
36:44
evolutionary context includes the
36:47
famous selfish gene notion of
36:49
Dawkins. Although I do
36:51
think of the many different sort
36:54
of lines of rationales and lines of
36:56
reasoning that Dawkins used
36:58
to support the selfish gene notion, the
37:01
one that I'm interested in is the one you
37:03
allude to, that is where
37:05
there's inter-genomic conflict, that it
37:08
makes most sense to think of a gene
37:10
as trying to achieve an outcome by
37:12
means of doing something that in this case
37:15
harms other genes within the same
37:17
genome. By now, a familiar
37:20
point and one that I think
37:22
Dawkins himself made in his
37:24
later work that the sense in which the
37:26
gene is the unit of selection where
37:29
there's inter-genomic conflict is really rather different from
37:31
the sense in which the gene is always
37:33
the unit of selection. I mean,
37:35
the latter is true in a sense,
37:38
but it's in a different sense that
37:40
we talk about genes as
37:43
the entity is being selected where
37:45
there's conflict within a single genome.
37:48
So I sometimes capture that by contrasting
37:50
the process of gene level
37:53
selection with a genetic
37:55
perspective on selection process that take
37:57
place at a different level. So,
38:00
Mir, can I ask you to tie back
38:02
genes as agents to your four types
38:05
of agents that we started with? Into
38:07
which category would genes fit for you?
38:10
Well, I would say really,
38:12
literally speaking, none of
38:14
them, because, I mean, you know, DNA is a
38:17
pretty inert molecule. I mean, obviously,
38:19
the more we know about how genes
38:21
work and, you know, how gene expression
38:24
works, you know, the more the
38:26
situation is a little bit complicated. But
38:28
I would still say that if we're
38:30
asking which of those agent concepts literally
38:33
characterise genes, I would be inclined to
38:35
say none. But nonetheless,
38:38
it can be useful to use one
38:40
or more of those concepts metaphorically
38:42
to describe genes for
38:45
certain projects. And I
38:47
think that, you know, what Dawkins and
38:50
Dennett and fellow travellers are doing is
38:52
basically taking one of those notions of
38:54
agency, namely the intentional agency one, and
38:57
using that notion, applying
38:59
that notion to genes, but in
39:01
an overtly, self consciously
39:03
metaphorical sense. I've always
39:05
been puzzled why people are so horrified by that.
39:08
I think it makes extremely good
39:10
sense to me. So long
39:12
as it's used carefully. Yeah, it's funny, I
39:15
feel relatively comfortable with, you know, the idea
39:17
of genes as a level of selection.
39:21
It's hard for me to conceive, and you
39:23
sort of said this, but it's hard for
39:25
me to conceive of genes as agents, because,
39:28
you know, to me, agency itself comes from
39:30
networks, doing computations and
39:33
making decisions. And of course, genes
39:35
don't contain those networks, right? They
39:37
can be a component of another network.
39:41
But they're just a part that is
39:44
participating in a greater whole that maybe
39:46
has agency, but that don't
39:49
have agency themselves. What do you think of that? Yeah,
39:51
I mean, I think I would agree. But
39:54
again, that's because you're using the terms
39:56
in this literal way. I mean,
39:58
I think that what say that
40:00
genes don't have themselves any
40:02
of the characteristic hallmarks of
40:05
agency, I think is true, but
40:07
I also think that that's compatible
40:10
with heuristically treating them as if
40:12
they did in order to further
40:14
our understanding of the evolutionary
40:16
processes, of certain evolutionary processes.
40:20
And so, I mean, if you
40:22
think of how people struggle to
40:24
understand the phenomenon of intra-genomic conflict
40:26
before they have the selfish gene
40:28
paradigm, which was really invented
40:30
for other reasons, really. Dawkins was
40:33
originally thinking about things like altruistic behavior
40:35
and so on. But, I
40:37
mean, I think, once you
40:39
have that idea, then
40:41
you can immediately see how to
40:43
make sense of this otherwise extremely
40:46
puzzling phenomenon. I think it'd be
40:48
useful to to give an example of
40:50
this type of inter-genomic conflict. Is there a
40:53
particular one that you really like? No
40:55
particular one. I mean, there are thousands
40:58
of these examples, but say, for
41:00
example, the phenomenon of cytoplasmic male
41:02
sterility in flowering plants,
41:05
where you have a gene within
41:07
a mitochondrion that causes
41:10
its plants to cease making
41:13
pollen, which is a really odd
41:15
thing to do. You know, the
41:17
plant, in some cases, the plant actually grows
41:19
stamen, so all
41:21
intents and balances in order to make
41:23
pollen, but then those
41:25
mitochondrial genes frustrate the
41:29
stamen from actually doing the thing that they're meant to do.
41:32
I mean, that's from the phenotypic perspective of the
41:34
whole organism. You look at them, what on earth
41:36
is that about? What's the
41:38
rationale for that? Is this an
41:40
exception to the theory of evolution or
41:42
something like that? How could it be?
41:45
But it makes perfect sense when you
41:47
realize that the mitochondrial genes are only
41:49
transmitted, but only down the
41:51
female line, so investing in pollen is a waste.
41:53
Now, that's the sort of case I have
41:56
in mind, where I say that there's a threat to the
41:58
idea that the whole organism is a waste of pollen. is
42:00
agent like because it
42:02
doesn't have if you like the
42:04
sort of phenotypic integration the behavioral
42:07
integration that one would
42:09
need in order to apply
42:11
the psychological language metaphorically
42:13
to it. So what
42:15
is it that the plant is trying to achieve? You
42:18
know is it trying to self fertilize
42:20
or not? Well on the
42:22
one hand yes because you know a lot of its
42:24
apparatus is geared to that but on the
42:26
other hand no because they don't make any problems. Yeah
42:29
and this this might be a bizarre thing to
42:31
say Samir to the author of a book on
42:33
levels of selection but you know those
42:35
kinds of arguments in the way you make
42:37
a completely fantastic case and there's no question
42:39
that these conflicts are true but
42:42
because biology is organized across
42:44
levels to sort of use
42:46
these as ways to understand
42:48
organisms you know is
42:50
it possible that these are great examples of genes
42:53
as parasites or something
42:55
different and the rest are a large
42:57
fraction of the genome and all its
42:59
regulatory components they aren't so
43:01
much parasites as commensals or mutualists I mean
43:04
you know to use the parasitic
43:06
argument to be foundational for everything
43:09
organismal I have always viewed that as an overreach.
43:11
Yeah I see I mean I think I
43:14
think I'm on board with the idea that
43:17
we can have genomic parasites and
43:19
that that actually accurately captures quite
43:22
a lot of cases of intragenomic conflict so
43:24
what's the thing that you think is overreach?
43:27
You know your arguments about why not
43:29
make pollen I mean that's sort of framed
43:32
in an adaptive context at the
43:34
organismal level where another explanation is
43:36
some type of frozen accident happened
43:38
in the past that produces this
43:40
lineage that you know the system
43:43
it's still alive natural selection is still going to
43:45
operate as natural selection
43:47
does as life does it makes the best of
43:49
what it has but to use that to
43:52
sort of expand that out to a full
43:54
way to understand the evolutionary process that
43:56
seems to be ambitious. Well you see what
43:58
I'm trying to do is to make an argument that
44:01
says, look, it's because phenomena
44:03
like that are relatively rare,
44:06
that the language of agency has
44:08
any purchase at all in
44:11
an evolutionary context. In
44:14
that we have this sort
44:17
of general philosophical line of
44:19
argument which says something like, look,
44:22
talk of intention, goal,
44:24
belief and desire is
44:27
really only appropriate when there is
44:29
a kind of rational integration indicative
44:33
of the fact that in cases, say,
44:35
of multiple personality then we really
44:37
struggle to say what on earth it is that a
44:39
person, you know, whether there really is one
44:41
person there at all or what it is that they believe
44:43
or what it is they want. We can't say it. You
44:46
know, the psychological language just ceases to
44:48
apply in that sort of case.
44:50
And what I'm arguing is that we have
44:53
a kind of biological analog of that in
44:55
the sense that if there were
44:57
too much intragenomic conflict, the
45:00
phenotypic consequences of
45:02
that would make it impossible for
45:04
us to think of the organism
45:07
as trying to achieve anything and
45:09
it would be impossible to metaphorically
45:11
characterize it as, you know,
45:13
aiming at some goal or other,
45:16
if you like. So I guess
45:18
I'm suggesting that this psychological integration
45:20
and rational integration is a necessary
45:22
condition of applying psychological language literally.
45:25
And the biological phenotypic integration
45:28
that one gets when intragenomic conflict
45:30
is effectively suppressed, as in the
45:32
vast majority of extant
45:35
organisms, is the necessary condition
45:38
of applying that psychological
45:40
language metaphorically to organisms.
45:43
And we see those types of conflicts not
45:45
only inter-genomically, but we
45:48
also see those conflicts, for example, between
45:51
the goals of males versus females and
45:53
how those play out, you know, in
45:55
terms of toxic sperm that may help
45:57
an individual male against set of conditions.
46:00
competitors but actually does harm to
46:02
the female. And so it's
46:04
the same conceptual framework that we're applying
46:06
in those cases as well. But in
46:08
that case, the conflict is between genes
46:10
in different organisms, I guess, or between
46:13
two separate organisms. Yeah. Or
46:15
yeah, two sexes, I guess. So
46:17
for that very reason, I think it would
46:20
be a mistake to think that a mating
46:22
pair was agent-like, if
46:24
you see what I mean, or to talk about what
46:26
the mating pair is trying to achieve. Obviously,
46:29
their interaction is partially mutualistic.
46:31
You use the term integration,
46:33
but also, I think
46:35
along with that, you also use the
46:37
term unity of purpose. And
46:40
so I think that's
46:42
the outcome of the integration, that
46:44
these actions are towards the goal,
46:46
towards this purpose, and it requires
46:48
this integration or coordination to achieve
46:51
that goal. Absolutely. But
46:54
I mean, one thing I do stress is that it's a matter
46:56
of degree. So there
46:58
are local violations of this unity of
47:01
purpose do certainly exist, and there are
47:04
relatively common in intra-genomic conflict, as you
47:06
guys all know, has been found in
47:08
almost every tax of where it's been
47:10
looked for. But nonetheless,
47:12
I do think that those
47:15
violations are necessarily local, if you
47:17
like, in that if you had
47:19
too much breakdown of unity
47:21
of purpose, you probably wouldn't
47:23
actually have an organism at all. And
47:26
you certainly wouldn't be able to think of
47:29
it as akin to a rational agent trying
47:31
to achieve an end to which its various
47:34
phenotypes all contribute. Yeah.
47:37
So I have a related question here that
47:39
builds on what you just said about the
47:41
idea of whether a mating pair
47:43
has agency, and that's just to
47:46
expand, it's also kind of a
47:48
levels question, but going in the other direction,
47:51
and thinking about groups of organisms and
47:53
whether or not they show agency and
47:55
how that relates to their evolutionary processes.
47:57
And like at some level, It's
48:00
easy to think of groups
48:02
that almost certainly do have
48:05
agency like U.S. Social Insect
48:07
Colonies in which many of the individuals are
48:09
closely related to one another and they act
48:12
essentially like a super organism.
48:15
But what about, say, interspecific
48:17
groups which occur
48:19
all the time in the wild? And I'm
48:21
thinking of things like, so there's a
48:24
retired professor here at Montana, Eric
48:26
Green, whose work some on communication
48:29
among different bird species in between
48:31
mammals and birds in forests. And
48:34
you may be aware of some of this work,
48:37
but there are waves of communication about predators
48:40
that travel between the members
48:42
of these interspecies groups. And
48:45
that really define a lot about where
48:47
these individuals go and, you know, when
48:49
they're feeding and when they're being fearful
48:51
and paying attention to predators
48:54
that are coming by. So would you
48:56
say that a large multi-species
48:58
group like that also shows agency
49:00
in some way? I
49:02
guess I would hesitate to say that,
49:05
although it might be true. I mean, I don't
49:07
think you could rule it out, definitely. But I
49:09
think the reason I would hesitate to say it
49:12
is just for the same reasons
49:14
that people have always been
49:16
reluctant to think of such a
49:18
multi-species group as an adapted unit
49:20
itself. I mean, clearly
49:22
the organisms in it are adapted and
49:25
in part they're adapting themselves to
49:28
the social environment that
49:30
they find themselves in, which includes
49:32
living in these multi-species communities. But
49:34
that's still different from saying that,
49:37
you know, the multi-species group itself is adapted.
49:39
I mean, this goes back to the point
49:41
that George Williams made way back in 1966
49:43
in his book Adaptation and Natural
49:46
Selection, where he
49:48
wanted to say, look, distinguish between a
49:51
group adaptation and a fortuitous
49:53
group benefit. We
49:55
can't really easily conceive of
49:58
an evolutionary process by which species
50:00
groups could have come to be
50:02
the adapted units themselves. Although,
50:04
I mean, it's possible. Because their
50:07
evolutionary interests are not closely enough
50:09
aligned consistently, right? Right,
50:11
exactly. Yeah, because we would think of
50:13
the evolutionary process as taking
50:16
place on the separate lineages
50:18
within those multi-species groups. And
50:21
to think of, you know, the multi-species group
50:23
as really part of the environment relevant
50:25
to the evolution of the, of
50:28
the populations, the
50:30
con-specific populations that
50:32
make them up. So, I mean, I think
50:34
that's the reason that I would hesitate to
50:37
apply the language of agency here. But
50:39
I would want to say that it
50:41
is ultimately an empirical question that, I
50:43
mean, there are people who, who take
50:45
seriously the idea of community level, natural
50:47
selection, where the community can
50:49
be a multi-species aggregate
50:52
of some sorts. But
50:54
I think if we sort of relax
50:57
the definition of integration, and I think
50:59
another way to capture integration is, is
51:02
to call it some degree of cooperation,
51:04
whether it's at the
51:06
whole organism level, you have to
51:08
have these coordinated responses. I
51:11
could also see cooperation, you know,
51:13
obviously within a group, but if
51:15
we kind of relax a little
51:17
bit what we, what we mean
51:19
by cooperation, I could see it
51:22
also applying to the examples that Art was
51:24
giving of, you know, mixed species,
51:26
because ultimately then
51:28
through that cooperation, what
51:31
emerges is this unity of purpose.
51:33
They all have the same interests
51:35
in avoiding the predators. Yeah,
51:37
I mean, I think that that's possible,
51:40
presumably that, that overlap of evolutionary
51:43
interests will be kind of
51:45
contingent, if you like, and
51:48
could break down. Especially if some of
51:50
those bird species are
51:52
also competitors and don't
51:54
have, under other
51:56
contexts, aligned differences. So yeah,
51:59
I think that kind of context
52:01
dependency probably plays an important role.
52:04
Right. But and of course, even if
52:06
the groups were not conspecific, you know,
52:08
comprised of a single bird
52:10
species, then I mean, it would still
52:12
be problematic to think of the
52:14
group of such a group as, as
52:17
an adapted unit automatically. I mean,
52:20
because I mean, I again, I would be happy
52:22
with that if a demonstrable
52:24
process of group level selection
52:26
has shaped the evolution of
52:28
some phenotype. So then
52:30
we get back into this question of, you know,
52:32
what sort of selection process is necessary
52:34
in order to produce an
52:36
adapted group, as opposed to
52:38
a group of adapted individuals, which
52:41
I've always thought is a helpful
52:43
framing that distinction that the Williams
52:45
made. Yeah. Do you
52:48
think that groups of humans, given
52:50
what we were just talking about with individuals within
52:53
a bird population, would groups of humans count as
52:55
a gentle or do you have to have the
52:57
same qualifiers for us? I
53:00
mean, I think potentially. So, I
53:04
mean, if you think, for example, of not
53:06
a group of humans, like a tribe
53:08
or a, but something more smaller, like
53:11
say committee, or something like that,
53:13
many people in philosophy of social
53:15
science take very seriously the idea that groupings
53:18
like committees can quite literally, literally
53:21
be agent like, we can talk
53:23
literally about what the committee believes
53:26
and decides and did. And
53:28
that certainly is manifested in the way we talk
53:30
ordinarily, we say, yeah, you know, the
53:32
firm raised prices, the committee decided what
53:35
to do, we do use this
53:37
language of choice and decision and
53:39
belief. And there's
53:41
a standard psychological language to
53:45
groupings of humans in
53:47
certain contexts. And
53:49
so I'm happy with the idea that that
53:51
could indeed be literally true. Yeah. And certainly
53:54
that it could be heuristically useful to think
53:56
that way. But again, only I mean, only
53:58
of course, under certain and circumstances. I mean,
54:00
if the committee couldn't come to a decision
54:02
or if, you know, one person on
54:05
the committee said, said X and the other said
54:07
not X, then of course we
54:09
couldn't sensibly talk about what the committee think
54:12
about the topic. So
54:25
we want to transition more directly
54:27
into the interplay between agency, your
54:29
ideas about agency and classic evolutionary
54:31
theory. So to put
54:33
this in an explicitly evolutionary perspective, and
54:36
we could stick with humans or we
54:38
can go in any direction you'd like,
54:41
is there evidence or do you expect that
54:45
in nature there is heritable variation
54:47
for agency? I think as
54:49
you were talking about agency there, it wasn't obviously
54:52
in the adaptive, the classic evolutionary adaptive
54:54
sense, but in natural
54:56
populations, what do you think about heritable
54:58
variation for agency? Well, I
55:01
suppose I would say, you know,
55:03
agency depends on, in any
55:06
of the senses that we've been discussing, it depends on
55:08
the possession of
55:10
certain phenotypes, such
55:12
as, you know, information processing,
55:15
sensing the environment, cognitive
55:18
abilities of various sorts. And
55:20
there must of course have been heritable variation
55:23
in all of those for them to evolve,
55:26
whether there still is now, is perhaps
55:28
slightly different question. But I still think
55:30
of that, I think of the heritable
55:32
variation as being with respect to phenotypes
55:34
that prerequisites
55:36
of or building blocks of
55:39
agency, rather than thinking of
55:41
agency itself as being something that
55:44
is itself a phenotype. Huh,
55:47
that's interesting. So you
55:49
wouldn't call agency itself a trait for
55:52
which there would be heritable variation? No,
55:54
I think it would, I think it's
55:56
almost at the wrong level itself to
55:58
be a trait. But
56:01
that raises a sort of an
56:03
interesting problem for me because, you
56:05
know, if we go back to Fischer, he
56:08
talked a lot about, in his
56:10
fundamental theorem, he talked about essentially
56:13
heritability for fitness. And
56:15
yet fitness is also made
56:17
up of all of these other components.
56:21
And so I see a lot of parallels there,
56:23
but yet in our models we don't have a
56:25
problem thinking about, you know, as long
56:27
as there's additive genetic variance for fitness.
56:30
And maybe that's because we're thinking about
56:32
it at a population level as opposed
56:34
to individual level. Yeah,
56:36
no, I mean, I think that's
56:38
an absolutely fair point. Yeah, so
56:40
if we take it that, you
56:42
know, fitness ultimately depends on the
56:44
vital rates, birth and the
56:47
death rates, and that
56:49
those that contribute to many, many different
56:51
phenotypes, of course, will contribute to the
56:54
vital rates of an organism of
56:56
any sort. So how is it
56:59
then that we're happy to talk about
57:01
heritability of fitness? Yeah, I
57:03
mean, I see the point. You're
57:05
saying, well, why by parody of argument, why
57:07
can't we talk about agency
57:10
as a heritable trait or
57:12
something exhibiting heritable variation? I
57:14
mean, in part, maybe the difference is in
57:16
part just that it's not so easy how you would
57:18
measure it. Yeah, I agree. Right.
57:22
Yeah, I mean, so do we find
57:24
an organism that's got more of it than
57:26
one that's got less of it? I
57:29
mean, how would you? Yeah, right. What does that
57:31
mean to see a population in which there's a distribution
57:33
of agencies, right? Yeah, no, that's
57:35
right. But nonetheless, I mean,
57:37
if we take it that in contemporary
57:39
biological populations, then we do see agency.
57:43
And presumably, if we go far enough back
57:45
in their evolutionary ancestors, then we don't
57:48
find agency or at
57:50
least only precursor forms of it, then some
57:52
evolutionary process must have taken place to produce
57:55
it. So in that sense,
57:57
it's got to be an evolved attribute. But
58:00
I'm still a little bit uncomfortable with thinking
58:02
of it itself as a trait, in
58:05
part just because it seems hard
58:07
to measure and to depend
58:09
on too many other things that have a
58:11
sort of clearer status as traits, let's say.
58:14
I agree, and I think that
58:16
actually is a point of argument
58:18
between me and Marty and Art,
58:20
that I recognize that organisms
58:23
behave and act
58:25
agentially, but I see
58:28
it more as sort of an
58:30
epiphenomenon. It's an emergent property that
58:32
probably has evolved over time, but
58:34
as an emergent property, it's something
58:36
that in of itself is really
58:38
difficult to, as Art was
58:41
saying, like, you know, plot
58:43
a distribution of agency in
58:45
a population. Absolutely.
58:47
I mean, I suppose someone might
58:49
make the argument, well, you
58:51
know, maybe it's fixed within
58:53
every species of something, so
58:56
it doesn't show any distribution, you know,
58:59
within a human population, for example.
59:01
But then it wouldn't be able to evolve. No,
59:04
but maybe it's evolved already. Maybe it's, you know,
59:06
it's like bipedalism or something. Interesting.
59:09
Well, it's a good thing in evolutionary biology, we
59:11
don't have any other great big emergent words like
59:13
gene and fitness and things like that. So,
59:18
Samir, you spent considerable
59:21
time or space in the
59:23
book focusing on Alan Grafven's
59:25
formal Darwinism project. And I
59:28
think the goal of Grafven is to
59:30
move away from population
59:32
measures of fitness to thinking
59:34
about individual optimization, in
59:37
particular as it relates, you know, to this
59:39
idea of agency. It
59:41
seems like Grafven's goals are
59:44
to separate selection
59:47
from individual optimization. But
59:50
then I think you made this
59:52
point in the book that natural selection may
59:54
not always then lead to
59:57
adaptation. And this
59:59
seems like a very counter... counterintuitive conclusion,
1:00:01
could you help me understand how
1:00:04
those get separated from one another?
1:00:07
How natural selection and adaptation get
1:00:09
separated from one another. Yeah, and
1:00:12
this idea that selection alone
1:00:14
can, in some cases, not
1:00:16
lead to adaptation. Right, yeah.
1:00:19
No, I mean, what I was
1:00:21
getting at with that was just
1:00:23
drawing attention to some points that
1:00:25
are fairly well known among evolution
1:00:27
theorists. Phenomena,
1:00:30
for example, are cyclical dynamics,
1:00:32
right, where even in simple one
1:00:35
locus population genetics then we
1:00:37
can point to certain
1:00:39
circumstances under which equilibrium
1:00:42
will not be attained, but
1:00:44
rather gene frequencies will cycle indefinitely,
1:00:48
never settling down and attaining equilibrium,
1:00:51
even though natural selection is the only force
1:00:53
at work, is the only evolutionary force at
1:00:55
work and it's not a stochastic effect. It's
1:00:58
nothing like that. So I
1:01:00
think simple examples like that seem
1:01:02
to illustrate, you know, fairly clearly,
1:01:04
I think that although we usually think
1:01:06
of natural selection as an
1:01:09
engine for adaptation, as producing
1:01:11
organisms that behave as
1:01:13
if they're trying to achieve some end,
1:01:16
some fitness related end, there's no sort
1:01:18
of logical guarantee that that's got to
1:01:20
be the case. And I think one
1:01:22
sees the same in models of frequency dependent
1:01:24
selection. So if you look in the popular
1:01:27
adaptive dynamics framework for thinking about
1:01:29
that, although the same is true
1:01:31
in other formal frameworks as well
1:01:33
for thinking about frequency dependent selection,
1:01:35
it's not always the case that
1:01:37
natural selection will take the population
1:01:39
to an equilibrium at
1:01:42
which individuals are playing
1:01:44
the optimal strategy conditional on the rest of
1:01:46
the population. I mean, often it
1:01:48
will. And that's the nice case in
1:01:50
which we get an evolutionary
1:01:53
stable strategy as the outcome
1:01:55
of frequency dependent selection. But
1:01:57
unpleasant, nasty, nonlinear.
1:02:00
dynamics are a cyclical dynamics for example and
1:02:02
other things are possible too as people have
1:02:04
known for a long long
1:02:06
time even though natural selection is
1:02:08
the only force at work.
1:02:11
So I think of I mean what Grafon is
1:02:13
doing as I see it is in part a
1:02:15
response to that situation in
1:02:18
that he's trying to say something like well I
1:02:21
mean can we say something about when
1:02:23
natural selection will be to adaptation and
1:02:25
he thinks that there's got to be
1:02:27
an intimate and tight link between them
1:02:29
and I think intuitively that makes sense.
1:02:31
I mean the whole point of the
1:02:33
theory of natural selection was precisely to
1:02:36
explain to phenomena adaptation and
1:02:38
diversity but then we face
1:02:40
this fact that when explicit evolutionary models
1:02:42
are made you know what obviously went
1:02:45
beyond what Darwin himself was
1:02:47
thinking of we find that
1:02:49
natural selection sometimes seems to result in
1:02:51
something that we can't really call adaptation
1:02:54
in any clear sense. So
1:02:56
then we want to say well can we
1:02:58
point out conditions under which
1:03:00
it will and that I think
1:03:03
of as what Grafon is trying to do
1:03:06
in his project and I think it is very
1:03:08
useful I mean although I don't think all the
1:03:10
moves he makes are ultimately successful
1:03:12
I think it's a conception it's extremely useful
1:03:14
to say look on the one hand we
1:03:16
have the process of natural
1:03:18
selection and on the other hand
1:03:20
we have the phenomenon of
1:03:23
adaptation to the environment and the
1:03:25
one is meant to explain the
1:03:27
other and the fitness concept rather
1:03:30
oddly is sort of features
1:03:32
on both sides of that equation. My
1:03:34
follow-up to that is it seems like
1:03:37
there's another concept here that is somehow
1:03:39
lost in the arguments which
1:03:41
is variation and so I
1:03:43
completely agree in the simplistic models
1:03:46
you know we think about selection
1:03:48
leading to adaptation and
1:03:51
the fixation of a single sort
1:03:53
of variant but whether it's
1:03:56
frequency dependence or density dependence or
1:03:58
just if there's fluctuation
1:04:00
environments, it seems like
1:04:03
one outcome of that is that
1:04:05
no single variant will always be
1:04:08
optimal. And a byproduct of that
1:04:10
is that variation is maintained in
1:04:12
the system, in the population in
1:04:14
this case. And so
1:04:17
I'm just kind of curious where
1:04:19
variation among different
1:04:22
agents or different individuals in a
1:04:24
population falls within
1:04:27
this idea. We touched on
1:04:29
this a little bit that
1:04:31
maybe bipedalism evolves and
1:04:33
then becomes fixed. But I feel like
1:04:36
at some level we
1:04:38
have to incorporate variation into
1:04:40
these ideas. Yeah, no, I
1:04:42
mean, absolutely. I mean, I
1:04:44
think everybody would agree that you need variation in
1:04:47
order for there to be natural selection at all.
1:04:50
If all organisms are the same in
1:04:52
some respect then there's no way that
1:04:54
some variants can out proliferate others. Everyone
1:04:56
has the same variants, if you like.
1:04:58
So that I think is presumably common
1:05:00
ground that variation is necessary
1:05:02
for natural selection. But
1:05:04
I mean, there's also a different question of
1:05:07
whether the outcome of
1:05:09
natural selection will be to
1:05:11
preserve variation or
1:05:13
to eliminate it. Yeah, we think about it
1:05:15
eliminating it usually. Right, we do. We do.
1:05:17
But of course, as you rightly point out,
1:05:19
I mean, if there's,
1:05:22
you know, environmental heterogeneity, temporal
1:05:24
or spatial, or a
1:05:27
host of other complications, or
1:05:30
non additive genetics, for example,
1:05:32
then we can have scenarios where
1:05:35
variation, genetic and phenotypic is in
1:05:38
fact preserved by natural selection or
1:05:40
natural selection can lead to adaptive
1:05:42
diversification, for example, as well.
1:05:45
So I sort of starting,
1:05:48
you know, assumption that natural selection will
1:05:50
reduce variation, although, you know, true, usually
1:05:53
for most, for the simplest
1:05:55
sort of natural selection need not always be
1:05:58
right. So. One
1:06:00
of the issues you brought up is that there's lots
1:06:02
of variation in the environment. And so often
1:06:04
there's not going to be a single genotype
1:06:07
that's optimal. Is
1:06:10
it possible that things like plasticity
1:06:12
and then maybe on top of
1:06:14
that, agential behaviors sort of writ
1:06:16
large are a response
1:06:19
to that and evolved outcome to deal
1:06:22
with that very kind of uncertainty,
1:06:24
right? So there's these systems that
1:06:26
are much more flexible over much
1:06:28
shorter time frames than alleles
1:06:31
and genomes are that
1:06:33
allow organisms to deal with
1:06:35
that kind of uncertainty. To deal
1:06:37
with environmental uncertainty. Yeah. Yeah. Yeah.
1:06:41
So you're thinking of an environment
1:06:43
that changes sort of within
1:06:46
the lifetime of an organism.
1:06:48
Yeah, within or between. But I
1:06:50
mean, either way, the systems that
1:06:53
produce these organisms during development inherently
1:06:55
have this sort of flexibility and
1:06:57
plasticity and agency built into them
1:06:59
to sort of compensate for the
1:07:01
fact that, of course, they're not going to have the ideal
1:07:03
alleles at all, the loci. There's no way
1:07:05
they possibly could, right? No,
1:07:07
I mean, I think that that may well be
1:07:09
true. I mean, of course, I
1:07:11
guess it's not the only possible way that national
1:07:14
selection could adapt a species
1:07:17
to environmental uncertainty. So I
1:07:19
mean, I think that's absolutely
1:07:22
right. But I mean, environmental
1:07:24
fluctuations and heterogeneity and uncertainty
1:07:26
are presumably pervasive in
1:07:29
the living world. I mean, pretty much
1:07:31
all lineages must face those. So then
1:07:33
I guess the question is why some
1:07:35
lineages evolved these agent like
1:07:38
behaviours or capacities
1:07:40
for dealing with them, whereas others
1:07:42
didn't. You know, so
1:07:45
an alternative might be to engage in
1:07:47
some sort of debt
1:07:49
hedging or diversification, for
1:07:51
example, where a given genotype produces lots
1:07:53
of different phenotypes in order
1:07:56
to hedge against the environmental
1:07:58
uncertainty. you know, if
1:08:01
one of those feed types will germinate,
1:08:03
whatever the weather turns out to be.
1:08:05
That well-known idea. So for me, the
1:08:07
question is, can
1:08:09
anyone say something sort
1:08:11
of concrete about the particular
1:08:14
sorts of environments that led to
1:08:16
the elements of agency being the
1:08:18
adaptive response as opposed
1:08:21
to sort of sub-agential things
1:08:23
that one finds more widely? Yeah.
1:08:25
Well, Samir, we've really enjoyed the conversation.
1:08:27
I really appreciate your effort. This has been
1:08:30
wonderful. Oh, it's been great. I mean, I
1:08:33
could talk about this stuff all day, as I'm sure you
1:08:35
guys could too. I
1:08:38
think we all could as well. Excellent.
1:08:41
So we always like to wrap with sort of
1:08:43
giving you a blank page to say anything about
1:08:45
the book or future plans or anything else you'd
1:08:47
like to talk about. Is there anything else you
1:08:50
wanted to hit? Not particularly. I'm currently
1:08:52
writing a book about the concept of
1:08:55
fitness and why it's caused so much
1:08:57
controversy and difficulty and why
1:09:00
it's not as simple a matter as it should
1:09:02
be. Okay. The subject for our next conversation,
1:09:04
please let us know. When that
1:09:06
one comes out, that would be wonderful. Okay.
1:09:09
I will. Hopefully this next year. Yeah.
1:09:11
Send us the book when it's out. Okay.
1:09:14
Well, thank you so much. Yeah. Thanks,
1:09:16
Samir. Thank you. My pleasure. Thank you
1:09:18
all. Thanks
1:09:36
for listening. If you like what you hear, let us know
1:09:38
by an ex Facebook, Instagram, or leave a
1:09:40
review wherever you get your podcast. And if
1:09:42
you don't, we'd love to know that too.
1:09:45
Right to us at info at big biology.org.
1:09:47
Thanks to Steve Lane, who manages the website
1:09:49
and Molly McGinn for producing the episode. Thanks
1:09:52
also to Dana Della Cruz for her
1:09:54
amazing social media. Katie Shumary
1:09:56
produces our awesome cover art. Thanks
1:09:58
also to the college of public health. University of
1:10:00
South Florida and the National Science Foundation
1:10:02
for support. Music on the
1:10:04
episode is from Poddington Bear and Tierra in
1:10:07
Costova.
Podchaser is the ultimate destination for podcast data, search, and discovery. Learn More