Select text to annotate, Click play in YouTube to begin
00:00:00
okay thanks tom um mike tom just introduced you there as a developmental biologist it always has seemed to me that you do developmental biology like no one else does um i suppose the the the general question
00:00:13
you seem to be uh trying to address is how cells organize themselves and each other in time and space to form sp uh to to form two distinct
00:00:24
tissues and organs and organisms with specific body shapes so really all the stuff between molecules and life as we observe it and it's in that in between space
00:00:37
that something seems to happen that distinguishes living matter from other kinds of matter um and i want to burrow down into that in what we talk about but first of all i wanted to ask you this my sense is that
00:00:50
you think there is something quite fundamental still missing from that picture of how we think about or at least how we talk about life something that we're not going to get at by just
00:01:02
filling in more of the details more of the fine molecular details so i want to ask am i writing that and briefly if i am what do you think that missing ingredient might be yeah well
00:01:13
i think that uh to to start to answer that let's let's go back to the first part of the question which is what what what do i think i'm really doing and i should say that uh i think that um developmental biology is uh which is a
00:01:26
uh an offshoot of what i think my fundamental interest is which is that i think what i'm actually studying is mind or cognition in different embodiments so so the reason that i'm so interested in developmental biology
00:01:38
specifically is that it's an amazing example where during the during uh sometimes just a few days and sometimes a few weeks you start off with something that people would call just physics or just chemistry so some sort of quiescent
00:01:51
oocyte it's a single cell it may not be doing very much and and after a very short time frame you then have something like an organism that has preferences ability to learn maybe it's if it's human eventually it
00:02:03
will have the ability to use language to expound on why it's not a machine and all of these are things that that people talk about and so that journey each one of us takes that journey right from from from from physics and chemistry to
00:02:14
complex mind and and and there's no there's no hiding from from that right to to whatever extent we are cognitive creatures we have an inner perspective we have various mental capacities we were once just physics all
00:02:27
of us were not just in an evolutionary sense but really in a developmental sense and you can watch it happen in front of your eyes so from that perspective i think developmental biology is is uh you know it's why i switched from doing computation in in sort of silicon medium to computation
00:02:40
and living media but i am fundamentally interested not just in questions of cells and why they do things but in morphogenesis or or pattern formation as an example of the appearance of mind from matter that's really right to me developmental biology is the most
00:02:53
magical process there is because it literally in front of your eyes takes you from from matter to mind you can see it happen and that's what i'm interested in and by that i understand i mean this is um
00:03:05
the the issue that uh tom mentioned of you your notion that and dan dennis notion that life is cognition all the way down and it seems clear you don't mean that as a metaphor do you you see a complete continuity between what single
00:03:18
cells do and what large brained animals like us do so where does that continuity come from what what what is your what what do you mean by cognition in this sense yeah so so so
00:03:31
the first the first point you know do i think it's a metaphor um you know in a strong sense i think everything scientists do is a metaphor so so i think we have to be really careful not to believe that we have some things that are sort of real pictures of the world
00:03:44
like pathways that we see in our textbooks i mean these things are are as metaphorical as anything else maybe more so and then there are these metaphors that you know we sort of we sort of speak about that aren't really real or just fuzzy ways of talking i don't
00:03:55
believe this i think that um in science all we have are metaphors now some metaphors are better than others and the way that i like to judge metaphors is to what extent do they facilitate progress meaning meaning new capabilities further
00:04:07
experiments right a good metaphor is one that it's not necessarily one that doesn't break prior assumptions it's not necessarily one that doesn't shock you it what it has to do is it has to uh move you towards new experiments and i think that
00:04:20
from that perspective uh cognition all the way down is is an absolutely excellent metaphor for a couple of reasons uh the first is that it's been incredibly fruitful in generating new new hypotheses new experiments new
00:04:31
capabilities um including some that are reaching into biomedicine we can we could talk about that but but the reason that the reason that i think it's it's inescapable is let's uh is is basically back to developmental biology instead of a lot
00:04:44
of people start off right you know they'll start off with with a paramecium or or you know some kind of simple life form and they'll say look surely that's not cognitive surely that's just physics right and and i i like to do it uh in reverse i like to
00:04:56
say okay what whatever humans have uncontroversially as high level cognition right so so if cognition means anything it it's we have it right otherwise it would just there's just no um nothing to talk about so whatever it is
00:05:09
that we have you can start with an adult human with all of this and just walk slowly backwards and when you walk slowly backwards you can do it evolutionarily too but let's just do it developmentally you walk slowly backwards and eventually you walk into a
00:05:22
a small child that's that's babbling and trying to understand the reference of language and before that you've got an embryo and before that you've got a single cell and before that you've got some chemical reactions and so the important thing is that biology offers
00:05:34
absolutely no bright line there is no process at which you can say that's where the lightning flash happened before that it was just chemistry and physics ah now we've got some cognition there is there is no no point like that that that anybody i mean people have
00:05:46
tried to draw lines like that i've never heard of a convincing one and so i think i think uh the way i would i would sort of flip that question around is to say if somebody if somebody doesn't think there's a continuity it's on them to
00:05:58
specify what the discontinuity is right it's not it's not it's it's not that you have to defend the continuity i mean darwin knew he spoke about this very clearly um the continuity is staring us in the face it's the it's the discontinuity that
00:06:10
needs defending if somebody in fact thinks that there is there is a sharp break here somewhere you just said something interesting you said you could look at that process either in evolutionary terms or in developmental terms um but i wonder
00:06:24
to what extent those are sort of equivalent ways of looking at it because developmentally there's um we can argue about whether the notion of a program makes sense developmentally
00:06:39
but certainly there seems to be a target of sorts there is something that you're going to end up with in evolutionary terms it's not at all obvious that that is the case and i mean it's normally said that it's not the case at all there's nothing
00:06:52
you know purposeful or directional perhaps about evolution but i wonder if that's something if that's a distinction that actually you want to challenge how did how should we think about that process as being distinct or similar in
00:07:05
evolutionary or developmental terms yeah um well i actually i actually have have written uh one one paper with chris fields on uh target morphology in evolution and we can we can talk about that specifically if people like but but
00:07:18
i want to go back to kind of um the the the really important thing that you just mentioned which is having a target of some sort one thing that uh is really important i i think one one one thing that we have to do is we have
00:07:30
to ask what do we mean by a cognitive system of some sort okay so when i say it's cognition all the way down what am i really saying so this has been this has been people people have addressed this for for a really long time and i
00:07:43
think we need to do two uh two interesting things one is that we have to abandon a binary view of things binary views get us in trouble in this in this case if we if if you try to say there are conscious systems and
00:07:54
unconscious systems there are cognitive systems and then there are just machines there are you know purposeful things and then there are artifacts those kind of um distinctions are completely arbitrary they are a result of of past limitations of
00:08:07
technology and imagination on our part they i don't believe they exist and they get us into all kinds of pseudo problems where people have to make these byzantine partitions and of things and try to figure out you know where things land what i think we have
00:08:19
is a continuum where one has to ask not whether something is cognitive but how much and what kind so there's a continuum so so i've lots of people ever since um cybernetics in the 40s where where norbert weiner
00:08:31
and colleagues have put out a an example of a spectrum like this i've got my own version which is called the uh the spectrum of persuadability that we can talk about but the idea is that it's a continuum and along this continuum there
00:08:44
are diverse capacities for goal-directed behavior the things on the left of the continuum have extremely small capacity for goal directed behavior so meaning that all they do is pursue energy gradients down and that's all they can
00:08:56
do these are balls rolling down a hill water flowing down and and things like this all the way on the right are very very clever things like humans which can pursue very complex goals using planning using extensive memory using all sorts
00:09:10
of neat tricks that they have in between this we have we have a variety of other agents that may be uh that may have limited limited ability to pursue various goals and i've the way i formalized it is as being able
00:09:23
to for any for any creature whether it be biological synthetic a software engineered alien you name it for any creature you can draw something called a cognitive light cone you can simply ask what is the size of the biggest goal
00:09:35
that that creature is able to pursue and when i mean biggest goal i mean distance in both space and time so for just to give you a simple example if you're a tick you've got a little bit of memory going backwards maybe a little bit of predictive capacity going forwards i
00:09:48
mean even yeast have that but really all you care about is your local concentration of butyrate then you're following the gradient and that's pretty much all you're ever going to do if you're a dog you have a much bigger horizon of memory going backwards you have more predictive capacity going
00:10:01
forwards but you're never going to care about what happens in the next town over three weeks from now just can't be done with that as far as we know with that cognitive system and if you're a human you might actually be depressed that in some billions of years the sun will burn out because you literally can care about
00:10:14
these enormous things very complex large scale things like global you know world peace and the status of the markets and all this so so what you can what the the thing about that kind of scale is that it takes all of this away from
00:10:26
philosophical uh sort of arguments and down into experiments basically we can actually find out we can do experiments what does that system what is it able to exert work towards what kind of states does it prefer and
00:10:38
how much how much competency does it exhibit for example and you know i forget who's um somebody had a great analogy and i forget who it was but this the difference between two magnets trying to get together and romeo and julia trying
00:10:52
to get together right the difference is how what what degree of of of ability to avoid local obstacles does the system have in doing it right it could be very simple it could be very very complex and then and then we can we can stop having
00:11:05
philosophical debates about this we can simply do experiments and that means that if if i show you some kind of a system maybe it's synthetic maybe it's a real animal who knows uh we can make hypotheses you can say i think the system lives at this level
00:11:18
of cognitive complexity and that's all and and somebody else might say oh you've missed the boat entirely in this other problem space it's way clever than you give a credit for here's my data here's my experiment and so then then we can we can do experiments and so now we
00:11:29
can talk about goals so uh goals in development um are uh what i've been what i've been working on for years is to try to show how the developmental system and this this includes development regeneration cancer
00:11:42
suppression all of these things is very very uh is has has large capacity for navigating that space of possible configuration we call it more for space it's the space if you're going to make a
00:11:54
head there are many different shapes of a head and there's this space of all possible configurations of the head so what we've been what we've been finding and people have been describing this for hundreds of years but i think not paying attention enough to this is that these
00:12:06
systems are incredibly capable of getting to the correct region of that space meaning making the correct shape despite all sorts of perturbations novel scenarios that they've never seen before evolutionary no evolutionary novelty i
00:12:19
mean i can give you many examples and so they have this amazing ability to navigate that space and so and so this is why i think of uh collections of cells as a meaning tissues during development or regeneration as
00:12:32
collective intelligence is navigating a space literally again no no no excess metaphors here literally we can port all of the math and the insights that we use to study collective swarm behavior and collective cognition
00:12:45
to understand how cells work together on these massive goals no individual cell knows what a finger is or how many fingers you're supposed to have but the collective absolutely does and the reason we know it does is because if you're a salamander and some of your
00:12:57
fingers are cut off what will it do it will regrow exactly the right number of fingers and then the most magical thing of all is it stops when it's done in order to stop when it's done it has to know what the correct shape is it has to know that the delta now is zero the
00:13:08
error is very low now we can stop so so in development there are absolutely goals like this uh in evolution i'm not sure but i'm not willing to say it doesn't exist and i think we need to do experiments i think it's unclear actually whether that's true or not
00:13:22
i i should say by the way for everyone listening please do post your questions which i hope you will have and i'm sure you will have at any point in the chat so that we're ready to go on to them um at any moment so don't you know don't
00:13:35
wait for us before you you pose them and then i'll i'll turn to you to ask them in turn um so mike it sounds as though there's um that there's a kind of a tension here between the notion in development that
00:13:48
there are these target morphologies and you talk about that in terms of your work on you know planaria for example where they find their way as you say reliably and reproducibly to these target morphologies
00:14:00
attention between that and the plasticity that we see in living form that there are alternatives and we can reprogram whole organisms and we can reprogram cells
00:14:12
how how does that balance play out you know to what extent do we have or does nature have the ability to take a different path and to what extent are those destinations predetermined yeah
00:14:24
yeah yeah great question um i i think it's not so much a tension as much as these are these two things are uh exactly um uh uh interdependent and and so and so uh the
00:14:36
thing the thing is that evolution is is not playing with a passive medium where it has to micromanage all of the details and and sort of figure out where everything needs to go evolution is
00:14:48
working with cells which themselves used to be independent organisms they have their own agendas they have um various computations that they can do they have preferences and so what evolution does is a kind of behavior shaping really
00:14:59
what evolution is really doing is pro is is providing uh various signals that get these complex creatures to meaning cells to do what it wants them to do evolution is hacking them the exact same way that bioengineers try to hack them in the
00:15:12
exact same way that cells hack each other in the exact same way that parasites hack other organisms this is all about figuring out um what is a level of control over the system that gets you the most bang for your buck and
00:15:25
what we've seen both as as workers in regenerative medicine and bioengineers uh and and uh just watching natural evolution the the most efficient way to control something as complex as a collection of
00:15:37
of of competent uh subunits like cells is not to try to micromanage it is actually to take advantage of the large-scale uh cognitive capacities that these things have as a very simple example you know what i what i usually
00:15:49
did say this to my students is look if you've got a rat and you want the rat to do a circus trick you've got two basic choices you can micromanage it meaning get in there and control every neuron like a play play it like a puppet and that will take you i don't know how many
00:16:02
years if that's even possible to to do that or you can just train the rack and this is why human beings have been able to train animals for thousands of years before knowing anything about neuroscience it's because because complex
00:16:13
systems expose interfaces including uh sensory systems including bioelectrics which is what we study that allow high-level controls it allow you to operate not the way that we used to program in the 50s where you physically
00:16:26
had to rewire things but and and the way that molecular medicine is current is certainly uh done today uh but actually uh at a much higher level by taking advantage of the intelligence of the system to operate at a higher level and
00:16:39
so what i think is going on here is that all of these system cells organs tissues and so on are navigating this more they're navigating these spaces now there are multiple spaces so there's anatomical space there's gene expression
00:16:50
space there's physiological space they're navigating all these things and uh they have certain policies for navigating them which include various preferences about what they want to do and various competencies things they know how to do and then things that
00:17:01
stump and flummox them completely and so absolutely there are multiple outcomes that can happen but all of that is uh by virtue of the way to understand that is not to try to uh try to build it up
00:17:15
at the very micro level but rather try to understand what what are the policies that are guiding these decisions and then how has evolution an engineer a parasite or again whatever it is of uh affected that landscape that now uh it
00:17:28
controls the way that that the system moves through that space so so just you know as a simple example is our our zenobots right so so what we do is we you know you ask um what does the frog genome know how to do well the frog
00:17:40
genome reliably uh makes a a tadpole that lives in a froggy environment and it's very fit for that environment and things are great well it turns out that that actually is so and so the skin cells of those animals are
00:17:53
a kind of a passive boring two-dimensional layer that sit on the outside and keep out the bacteria and you might think well that's what skin cells know how to do and that's that well it turns out that if you actually liberate those skin cells from the rest of the animal and put them on their own
00:18:06
and give them a chance to re sort of reboot their multicellularity what you find out is that their normal two-dimensional passive life is just uh the result of instructive interactions or or or re really a kind of uh behavior
00:18:18
shaping or or bullying by the other cells that's what they do when the other cells are making them do this on their own they do something completely different they form a little creature known as a zenobot it has autonomous motion it has lots of other behaviors including the ability to make new
00:18:31
zenobots by running around and collecting loose cells which it then packs into the new generation of zinabots and it's amazing and that kind of thing has never existed uh in nature to our knowledge before it's certainly never been part of the frog life cycle
00:18:43
it's it's a different way of being it's a different um uh set of attractors in the space of possible morphologies and behaviors and those cells are able to find that that's that's that's massive plasticity and yet
00:18:55
in their normal environment they will make a very good tadpole and a very good frog and in fact if you scramble the craniofacial organs of that frog of that tadpole they'll you'll still get a good frog because all the organs will move around relative to each other and land
00:19:07
in the correct orientation and then stop right so so i think this this ability to um that that's the mark of a of a sophisticated navigational system is that is that you you can get around this space but it and and you'll get to where
00:19:20
you're where you're going despite perturbations but at some point you might need to go somewhere else and that's and these systems can absolutely do it i think i think that kind of thing potentiates evolution greatly evolution has the speed that it does exactly because of that
00:19:32
fantastic
End of transcript