FCE Conference // David Chalmers, “What is Conceptual Engineering?”


Many thanks, as you can already tell, Vera
really did all the work in organizing this conference, from the conception
of the whole thing to the execution, so we’re all very grateful
to Vera–except for the part where you put me on the program at 9 o’clock in the morning.
[Audience laughs] But I’ll do my best.
So my original thought was to call this talk “Conceptually
Engineering Conceptual Engineering” but I found that title was already taken by
Manuel Gustavo Isaac. Where’s Manuel? Yeah, terrific paper, I recommend it to
you, but taken. So I couldn’t use that, so instead I went for a title to
pay tribute to the Sally Haslanger mode of conceptual engineering. What is
conceptual engineering, and what should it be? My original thought was I was
gonna give a talk with a bunch of theses about conceptual engineering, where the
only example and the only case study was gonna be the concept of conceptual
engineering. But that got a little bit too vertiginous.
So we’ll have a little bit of that in the talk. But there’s gonna be a bunch of thoughts
on different kinds of conceptual engineering, and some of my
thoughts about what some of the most important and perhaps
overlooked kinds are, and how they play into the importance of conceptual
engineering within philosophy. So I’ll start by talking about kinds of conceptual
engineering and different varieties, and go on to issues about the importance and the
difficulty. So what is conceptual engineering? There is a very obvious way to come at
this. To find the definition of conceptual engineering, go look up the definition of engineering,
and then just appeal to compositionality. [unclear] concepts in wherever possible. I found a lot of definitions of engineering
on the web, by different engineering associations. The mechanical engineers, and
the civil engineers, and the electronic engineers. There was a very simple one, which, actually, a lot of
the more complicated ones are basically just, you know, variations on–making variations
on this theme. You can find this one quite commonly: the process
of utilizing knowledge and principles to design, build, and analyze objects. Now,
you’ll notice–you’ll start from those three. I mean the key thing there is the design,
build, and analyze. Variations on those three you’ll find in most
definitions of engineering, some of them extend that list to about 20 different
things. There’s like, to operate, and to maintain, and to repair, and
to forecast, and to evaluate. But they basically all, in some broad sense, fit within
this rubric of designing, building, and analyzing. Okay, well, then
just invoking compositionality, conceptual engineering will be something
like the process of utilizing knowledge and principles to design, build, and analyze
concepts. It’s not bad, except ‘analyze concepts’ already has a meaning, which is maybe not totally
a propos in this context. And building isn’t obviously the right word. With a little
tweak, we get, the process of utilizing knowledge and principles to design, implement, and
evaluate concepts. Something like that I think is not a bad first pass
at what conceptual engineering is all about. And that then gives us
different broad stages or types of conceptual engineering.
There’s the design stage, where we design concepts. Various ways to do that. One
classic way would be to give a definition, or maybe an
inferential role, or some paradigm cases, or something like that. That’s the design
stage. Next is the implementation stage, a bit like the, you know, actually building the
bridge, and so on. In the implementation stage you actually have to use a concept,
and maybe try to get others to use it too. This is, I think, what Herman calls conceptual
activism. And then there’s the key evaluation stage, which play a central role in the
conceptual ethics work, so people like Alexi and David, where what’s key is, you know,
the evaluation of how good these concepts are in themselves and for certain
purposes. For example, to see how well they play key roles. I mean, you can see
all three of these things playing a role, say, in bridge engineering.
You design a bridge, you implement a bridge, you evaluate the bridge to see how
well it’s doing. And I think you see something like this in software engineering.
Design a program, build a program, evaluate the program. I quite like
the software engineering analogy because, you know, many people
worry about, how does conceptual engineering work if concepts are abstract objects?
You know, you can’t exactly build a new–it’s not obvious, at least, that you can build a new resident of
the third realm in this way. Exactly the same issues arise for software
engineering. You know, programs are arguably abstract objects, and there’s
complicated things but fairly obvious things you can say in both cases about
what goes into the implementation stage, or maybe what’s relevant to building and
implementing a new thing is a matter of coming to stand in new relations to
those things. Anyway, I’m not gonna make heavy weather of abstract objects versus
concreta here, ’cause I think it’s fairly natural to– [unclear]. Now, another
relevant distinction here is creating versus fixing. Designing and implementing
a new, say, bridge, from scratch, or fixing an old bridge. This is very relevant in
ordinary engineering. Very relevant in software engineering, and I
think ought to be very relevant in conceptual engineering. And this is
really going to be what I’m focusing on to a considerable extent. But first I want
to just go through a bunch of examples of what I think of as paradigmatic
conceptual engineering, in both modes: the creating and the fixing mode. All within
philosophy. Not because philosophy is the only locus of conceptual engineering.
I think it’s absolutely everywhere. Philosophy is the one that I know best, and that many
of us here have particular expertise on, so we might as well talk about what we know. So take something from metaphysics. I
think, for example, the concept of supervenience to me is a paradigm
example of conceptual engineering. Someone said it has the smell of
something that was thought up in the metaphysics lab.
[Audience laughs] This idea of these properties,
which depend on some other class of properties. You duplicate one, you duplicate the
other. This concept, I think, was engineered over the–Moore had it without the name, Hare
introduced the name, Davidson and Kim and others made much of it. I think that ought
to be regarded–to me that’s paradigmatic conceptual engineering. And indeed the
notion of supervenience was once thought to be one that could do a lot of
philosophical work that previous concepts like identity might have been
hoped to do as well. Then later on people thought, okay, that doesn’t do that work quite so well, so
now we’ve got these newfangled concepts like grounding. A bit more conceptual
engineering or at least conceptual abstraction of ideas, which are in the
air, under a new label. So these are examples in the mode of creation. In
the mode of fixing I think, for example, of Amy Thomasson’s work on existence in
metaphysics, where taking a very pragmatic approach to what is the notion of, say,
existence or of object, which will serve us best. I think conceptual engineering in this
mode is is everywhere throughout the philosophy of language. Semantic values
and notions meaning are engineered all the time. Carnap’s notion of intention,
Frege’s notion of sense, Grice’s notion of implicature, Kripke’s notion of
rigid designator. I think all of those can be seen as, you know, engineering
fruitful new concepts in the philosophy of language. In the revisionary mode,
there’s various work on truth. Actually Carnap’s work on explication
took truth as one of his central examples, and he saw Tarski’s explication
as a kind of conceptual engineering. More recently
Kevin Scharp diagnosing truth as an inconsistent concept, and
proposing his upward and downward replacements. In philosophy of mind–I was
saying in our philosophy of mind seminar the other day that professor Block is a
paradigm conceptual engineer. His 1995 paper on the function of consciousness
says consciousness is a mongrel concept, problematic in all these ways, let me
find you these more precise and interesting concepts, such as the concept
of access consciousness, engineered and useful. Herman in his book on
conceptual engineering puts forward my paper with Andy Clark, “On the Extended
Mind”, as a paradigm case of conceptual engineering, in engineering the concept of belief. My attitude towards that is slightly complicated, and I’ll
get back to that in a moment. Tamar Gendler on ‘alief’ is another person–is someone else
that is engineering a new concept in the vicinity of belief. In social philosophy
you find this kind of thing all the time, I think. People pointing to
phenomena that may have been overlooked with useful concepts, or drawing
distinctive concepts out of strands of discussion. So I think
Miranda Fricker, epistemic injustice and its varieties like testimonial and hermeneutic
injustice, would be a paradigmatic example here of drawing out a
fruitful concept. Sally Haslanger’s work would be–on gender and race. The paradigm
example would be her work towards the analysis of the concept of woman in terms of
oppression. Sally calls it ameliorative analysis, that’s conceptual
engineering in the revisionary mode, and this ameliorative strand of conceptual
engineering has been picked up by many other people in recent social philosophy,
like, for example, Kate Manne’s revisionary analysis of misogyny. In metaphilosophy,
indeed, the field that is roughly the field of
this conference, you can find a few paradigmatic examples. I think Carnap on
explication was itself a wonderful example of relatively de novo
conceptual engineering. Here is a new and useful concept. And indeed the very
concept, you knew it was coming, of conceptual engineering, I think, is a marvelous
piece of of conceptual engineering. By the way, I’ve noticed in the literature on
conceptual engineering, the credit for the phrase ‘conceptual engineering’ seems
to be going largely to Simon Blackburn in 1999. If there is one thing you remember
from this talk, it’s not Simon Blackbird, it’s Richard Creath 1990,
in the literature on Carnap really made a big deal of Carnap as a
conceptual engineer. It first plays a role in his book Dear Carnap, Dear Van, and
many other papers. Conceptual engineering has been
all over the Carnap literature for decades. PhD theses have been
written on it. So I think it would be nice for Creathe to get some credit here. Carnap
himself actually talks about linguistic engineering in a couple of places, but
never uses the phrase conceptual engineering. Thanks to Manuel Gustavo
for some discussion of that. In any case, I think,
this concept of conceptual engineering is itself a model of a useful bit of
conceptual engineering. So okay. So now with those examples in place, let’s think
about the whole fixing versus creating distinction. Herman’s definition, I think,
of conceptual engineering basically ties it very centrally to the fixing
project. The title of the book is Fixing Language. I think the more detailed definition is
assessing and improving our deficient representational devices. So maybe this
sounds like–okay, that’s part of conceptual engineering. If we said
that civil engineering, say bridge engineering, was the project of fixing
and improving our deficient bridge devices, or that software engineering was
fixing and improving our deficient programs–well, that’s part software engineering, bridge
engineering. But, between us, not necessarily the most important, not necessarily the most
exciting part. It’s one very important part. There’s
also the whole project of building new bridges and building new software. I think
Herman in his book acknowledges the possibility of this project, and goes on
to say something like, that’s just not the one I’m interested in here, but I
do think it’s worthy of our attention. So I encourage making this
distinction, between what I call de novo engineering and re-engineering,
where de novo engineering is building a new bridge, program, concept, whatever. And
re-engineering is fixing or replacing an old bridge, program, concept, or
whatever. The name is still up for grabs. At one point I was using de novo versus de
vetero, and someone pointed out to me that wasn’t really proper Latin.
[Audience laughs] Anyway, so I’m open to ideas.
It’s not totally straightforward to draw the distinction. There are some hard
cases. Here’s the Tappan Zee Bridge, just up the Hudson River from
here, and here’s the old bridge, the old Tappan Zee Bridge, here is the new Tappan Zee
Bridge. And the old one is still there, and then they’re basically building a
new bridge in the same location as the old bridge, in order to replace the old
bridge. Is that re-engineering or is that de novo because it’s a new bridge, or
is it re-engineering because it’s a replacement. I guess for my purposes
I’m gonna count that kind of thing as re-engineering, because the central theme
is it’s being used to somehow to repair or fix an old bridge, and at the end of
this process the old bridge will be gone. Likewise for de novo–for conceptual
re-engineering, it’s not like you keep all the concepts around afterwards.
The point is fix this very concept. We can argue about how we draw the
lines. And many of the standard examples in the
conceptual engineering literature are conceptual re-engineering.
Certainly–you know, the Carnapian explication literature’s very much a
literature on re-engineering. And Georg Brun has a nice paper, actually, on explication
as conceptual re-engineering–he uses that exact phrase. Belief has
been one of the central candidates for
re-engineering since the 40s and 50s–it certainly didn’t start with Andy and me–say, with
various probabilistic analyses. The concept of truth, more recently social
concepts, the concept of woman, the concept of race. Many of the examples I gave, on the other
hand, I think, look more like de novo conceptual engineering: epistemic injustice,
supervenience, rigid designator, and, indeed, conceptual engineering. These weren’t
particularly trying to fix or replace other concepts. If you squint really
hard, you say, oh, maybe supervenience is intended as a replacement for identity. But
actually, not really. Identity’s doing fine, it’s a great concept. This is one job that identity people were
using for to do some reductive projects that people try to use supervenience to do. Or
maybe conceptual engineering is a replacement for explication. I don’t
know. I mean I think it’s not the most productive way to think about these
things as re-engineerings. Here are some useful, fruitful concepts that we can use
to do some interesting philosophy with, and that may have useful consequences.
Anyway, so my proposal is that conceptual engineering either includes
or should include–that’s a little bit of weaseling–both de novo conceptual
engineering and conceptual re-engineering. So you might be inclined to
think the should version of this suggests this is at least potentially a proposal
for conceptually re-engineering conceptual engineering, in which case, you
know, conceptual re-engineering is important too, and this would be an example of
it. Because that would apply only if conceptual engineering doesn’t already
cover both de novo conceptual engineering and re-engineering. My own
view is that conceptual engineering already covers both. Partly in virtue
of compositionality. I think it’s just, you know, it’d be very weird if conceptual
engineering worked in such a different way from other kinds of
of engineering. Partly in virtue of the unity of the category, partly
in virtue of other factors. Hey, I may be wrong. You know, maybe the use of
conceptual engineering by various people in this community recently focusing on
re-engineering has given us some semantic glue that makes it stick to
that, so, you know, I don’t want to put too much weight on the semantic claim,
although I think I’m right. But in any case, it’s a verbal dispute about
conceptual engineering. Which actually leads to my sideline on conceptually
engineering belief. What I said about my attitudes towards conceptual engineering, I
think, was Andy’s and my attitude to the conceptual engineering of belief back in
“The Extended Mind”. What we said and Herman quotes: “We don’t intend to
debate what is standard usage; our broader point is that the notion of belief ought
to be used so that Otto qualifies as having the belief in question.” Where Otto
is the guy who carries around the notebook which his memories are stored in. I think
actually we didn’t–I mean, I’m very happy to be taken as a paradigm of conceptual engineering,
I’m a fan of conceptual engineering, but I’m not sure that we actually saw this
as conceptual engineering. Our own view, I think, was that these extended cases of
beliefs were literally beliefs. So the word ‘belief’ already covers them; perhaps in
virtue of unity and the explanatory fecundity of the category.
So that ‘believe’ would already include these extended cases. We were just saying, if someone wants to argue with that, here is the underlying more important claim. I
think I’m inclined to say something about that–about this claim about
conceptual engineering too. I think it does cover de novo conceptual engineering. If
you think that’s not the actual meaning of the term, then, well,
I think it should be used to cover de novo conceptual engineering.
And indeed in other cases, my philosophical ideology is, you
shouldn’t get too hung up on words. You could have used–introduced a new term,
‘e-believe’, to cover all these cases, and just made all these
claims about how unified it is with the ordinary cases of believing, and said this
plays the most important role. Yeah, we could have done that, but what fun
would that have been? You know, it’s like the word ‘belief’ has got certain–
you know it’s used a lot, it’s got certain attractions in explanation,
attaching the word ‘belief’ to a project, I think, it’s got certain pragmatically
useful roles. Likewise the word ‘conceptual engineering’. Conceptual
engineering is cool, people have conferences on it. So therefore–it’s got
what Herman calls a ‘lexical effect’. It’s like, yeah, calling or hearing another name,
like de novo this and that–that’s not going to work in the same way. So
pragmatically it makes sense to try and attach this thing you’re
interested in to this word. Is this a verbal dispute over conceptual
engineering? Yes, but a nonverbal point: conceptual engineering should cover de novo
conceptual engineering, because of this compositionality and this unity, and
because de novo conceptual engineering is at least, I think, as
important in philosophy and elsewhere, as conceptual re-engineering.
This immediately leads us to delicate questions about the connection between
conceptual engineering and linguistic engineering; sometimes these phrases are
used almost interchangeably. Certainly it’s the case that wherever you find
conceptual engineering, you almost always find some linguistic engineering.
Whenever there’s a proposal about a new concept, there’s also a proposal about a word
for it to be attached to. I mean, this is actually–it’s a version of what Peirce
calls “the ethics of terminology” that Steve Yablo turned me on to years ago.
And all this really kind of comes down to the ethics of terminology,
or the ethics of language. I call it linguistic ethics in the
current modes–how we should be using new and old words to express
new and old concepts. So in the linguistic mode, I think there’s a
related distinction, which is the distinction between homonymous and
heteronymous conceptual engineering. Same-word linguistic engineering–fixing
meaning for an old expression, and maybe the cases like ‘woman’ or ‘misogyny’ and
‘truth’ are homonymous. I think homonymous conceptual engineering–at least the
clear cases–are always best seen as conceptual re-engineering: fixing the old
concept expressed by the word. I mean there will be cases where you can use an old
word for something totally new: supervenience did already exist, as
a way of meaning something else. I don’t think it’s usefully regarded, though, as–it
wasn’t as if you’re particularly fixing the meaning of that word; in that case
it’s more or less coincidental that that word is being used. Heteronymous
conceptual engineering is different-word linguistic engineering.
Re-engineering an old concept sometimes–like Ned re-engineered the concept
consciousness to the concept of access consciousness. Instead of saying,
let’s use the old word to express consciousness–that would be homonymous–let’s
introduce a prefix–very very common strategy in philosophy–access
consciousness, to express the new concept. So that makes it now heteronymous. Or in
other cases a wholly new concept: supervenience and, indeed, something like
conceptual engineering may be like this. Again, the existing literature focuses almost
entirely on the homonymous case, but I think the heteronymous case is just as
interesting. One reason why it’s–one thing that makes it interesting in this
context–I think it’s at least somewhat less vulnerable to challenges such as
Strawson’s challenge: aren’t you just changing the topic? To which the answer for
heteronymous conceptual engineering is, sure, I’m changing the
topic–here’s an interesting topic. And indeed the
externalist challenge–it’s not to say that worries about externalism don’t arise in
coining a new word. But in the use of an old word, they really arise big time,
because there’s that whole social community of users of the word to defer to,
making it all the harder to affect conceptual change.
I mean, when it comes to linguistic engineering, I’m sort of on record in my
paper on verbal disputes as saying, the words we use for these things really
doesn’t matter all that much in philosophy. So that suggests a kind of
critique here of the homonymous, and indeed the heteronymous, project. For substantive
issues, it doesn’t matter what words we use. Why does it matter whether we use a
new word or an old word to make our claim? So what’s the point of homonymous
conceptual engineering or indeed of heteronymous
conceptual engineering? But I think we’ve got to
distinguish here between broadly theoretical and broadly practical
projects. For theoretical purposes, I think in principle anything you can say
with homonymous conceptual engineering can also be said with heteronymous
conceptual engineering, and vice-versa. You can use an old word, you can use a new word.
For theoretical purposes it’s just a verbal difference between those two. And indeed, as
I say in the verbal disputes paper, for an ideal reasoner this difference
wouldn’t matter. But that’s it. We aren’t ideal reasoners. We have
strong associations with the use of certain words.
I mean, projects in heteronymous conceptual engineering could be done homonymously. Instead
of introducing the new word ‘supervenience’, we could say, here is what
reduction should mean. And then give a supervenience definition. Instead
of using the word ‘sense’ let me appropriate the word, let’s say, ‘meaning’.
Things could be done that way. And vice-versa for various homonymous
projects, like the revisionary analysis of truth–I think
Kevin Scharp does end up using a different phrase, ‘upward truth’.
For Sally’s ameliorative analysis of ‘woman’–I mean, there’s certainly gonna be
a related project that says, let’s use a new word, like ‘womyn’ with
a y, to to express it. But I mean the practical effects will be different.
The upsides and downsides here are all– probably all–practical rather than
theoretical. The lines can blur in some cases. Sometimes we can achieve our
purposes better with a new word, and sometimes with an old word. I think one
of the papers tomorrow, is it Ari? He’s giving a paper on the on the upsides
and downsides of using the same words and different words.
So just a couple of initial thoughts about this. Well, there’s
entrenchment: new words are expensive and harder to get people to use; old words
have associated prestige, like ‘conceptual engineering’–so prestigious!
People having a conference on it. Makes sense, if you want to get your concept expressed, attach it to
that powerful engine. Role: the old word has certain fixed roles.
Homonymous conceptual engineering can allow a new concept to very easily become
associated with those roles. Certainly, I think, in the case
of, say, gender concepts like woman, look at all the roles that plays in
society; you want something different to play those roles, and this is the way to
do it. And this ties very closely to projects in, say, social
justice. Given fixed social roles for terms like ‘woman’ or ‘marriage’, homonymous conceptual engineering allowing a new concept to
play those social roles, can in many cases made for a more just world. So those
are some of the upsides. I mean there are downsides to–homonymous
conceptual engineering, especially for theoretical purposes, can be very
confusing, with all these multiple meanings floating around,
extremely difficult to implement–that’s something which Herman has focused on.
Unless one is very powerful or very lucky– you know, maybe in a small
community: there’s only thirty people in the world working
on conceptual engineering and they’re all in the room, so you have a chance. But
even then it’s not a big chance. Heteronymous conceptual engineering
is perhaps somewhat easier to implement, at
least in limited circles, to put forward a new word. Those are upsides and
downsides. Let me just say something quickly about the importance of
conceptual engineering– some people say it’s all of philosophy.
Is it important? Yes, some of the most important advances in philosophy quite
clearly involved conceptual engineering, especially some of those cases I gave
of de novo and heteronymous conceptual engineering. You’re making new
concepts to do a lot of work. Is conceptual re-engineering important in philosophy?
It’s very practically important, especially for roles outside philosophy, and
philosophers have a role–a central role to play, I think, in that project. Still,
theoretically, within the philosophical community, you know, here and there
there are cases where it’s played a useful role–cleaning up a concept and
giving a nice analysis. Maybe probability and the different notions of probability is
one example of that. I think most of the time my view is it can be done just as well in
heteronymous or de novo mode. We’ll come back to the reasons for
that in just a moment. Is conceptual engineering everything in philosophy? This is
clearly false. Lots of important philosophy involves arguing for theses
using old languages, and is none the worse for that. Think case like Jackson’s
knowledge argument or Parfit’s repugnant conclusion. As far as I can
tell, there’s no real conceptual engineering in any of these cases; you can
use the old concepts still for very interesting phenomena and theses.
Flocke: Dave, five minutes. Chalmers: Okay, good. Even when philosophy
involves new concepts, I think it typically also involves new theses involving those
concepts. I think the role of theses here is very very important–enough to be
underestimated. So why is supervenience interesting? There’s no way you look at the concept
of supervenience and say, wow, how cool a concept is that. Rather, it serves
some potential roles for you that can connect to certain theses. Like many
people think, for example, that physicalism requires supervenience. Or perhaps more strongly,
some people think if you have supervenience, then you get physicalism.
Supervenience suffices for various reductive projects, and then, you know,
someone like, say, Kit Fine comes along and argues with that, and we can
assess those theses. Or implicatures. It’s like, ok, implicatures are
interesting, but really where the rubber meets the road, or you know, will these
maxims govern implicatures? Implicatures play these roles in
discursive practice, and so on. So really it might be the theses are where the rubber
meets the road. The theses have a certain primacy in philosophy, at least the
sort of philosophy aimed at discovering truths. This is not saying that concepts
are not important. It’s to say that somehow the importance of concepts, in my view,
to a very large extent derives from the importance of theses that they’re
involved in. This is something that Carnap tacitly recognized in his
discussion of explication, where he talked about the fruitfulness of a concept,
as being on one of the dimensions of assessment. The fruitfulness of
a concept comes down to the number of interesting and useful and
explanatory theses it might be involved in. So on this view the importance of
concepts derive from the importance of theses, and certainly formulating theses
and arguing for them goes well beyond conceptual engineering, on my view.
And largely, I think, to a very considerable extent, we should see theses
as driving conceptual engineering. So this whole debate, grounding
and supervenience, which concept is best in metaphysics, really comes down to, you
know, what are they useful for? What theses can they be used in, for
example, for reductive projects like physicalism, is grounding or
supervenience the best conditional. For assessing the status of various concepts
of meaning, let’s look at the different properties of language that they can
explain, theses you can get them involved in. So this leads to the
metaphilosophical view of these things which I favor, concept pluralism, which Herman
also cites in this book as a kind of conceptual engineering view. There are
many concepts in the vicinity of most philosophical words playing different
roles; we ought to articulate those roles and find the concepts that play
them best. I think that’s a kind of conceptual engineering: finding
concepts that play these roles the best, we’re evaluating concepts in many cases,
finding relatively new concepts that can play them. That’s a kind of conceptual
engineering. For the most part I don’t care whether it’s homonymous or
heteronymous, except for these practical purposes. But again, it’s not just conceptual
engineering. It’s articulating theses of the form ‘x plays role y’. You can get a cool
concept when you get some cool theses. For a concept–I think–this
maybe is the strongest thing I’m going to say in favor of de novo conceptual
engineering. For a concept pluralist, de novo conceptual engineering is often better
than re-engineering. ‘Cause basically, if you’re a concept pluralist, I think it’s very often the
case that–my paradigm of conceptual re-engineering is, you somehow fix
or replace the old concept, so the old concept is no longer around. For the
concept pluralist, a lot of time you want to say, why not have both? Even, say, for
the case of Sally’s concept of–analysis of woman. Even if you think that somehow
that the old concept–the old, you know, say, biologically-based concept of
woman was an unjust concept for various social purposes. There’s all kinds of socially
unjust effects. Somebody might still think, well, nonetheless it’s a useful concept to have
around, say for certain medical purposes, and one doesn’t have to–just because a
concept is useless for some roles, doesn’t mean it’s useless for all
roles. So sometimes I think, in some cases, the old concept is so defective, maybe an
inconsistent concept, an imprecise concept, an immoral concept, that it’s better gone.
My own view is concepts which are that defective are actually
fairly rare, and it’s not the the typical case. We can argue about that.
But in cases where the old concept is not defective, the concept pluralist says, why not
keep them both around for different purposes. Then we have a very critically fruitful
apparatus. I was going to say something about the difficulty of conceptual
engineering. I think basically my view is that it’s possible, clearly. It is also
very clearly difficult; something that Herman stresses, by the externalist
challenge. I mean, it’s a very difficult social project. I think that maybe
breaking it into stages helps. We divide conceptual engineering
into (i) designing a concept, such as proposing a meaning for a word, and (ii)
implementing the concept, say ensuring the word is used to express this concept
within a community. The design project is not that hard. You propose a definition
or inferential role or application to paradigm cases. It’s not clear that externalism is a major
obstacle to that. Implementing it is harder work. You gotta use it that way, gotta get
others to use it that way. That’s hard. And then, even after it’s used that way, you gotta ensure
that the meaning, the reference, is right. First–even the first is non-trivial, the
second is highly non-trivial: getting others to use it that way.
Externalism I think of as mainly affecting the third step. The step from
use to meaning. The externalist is gonna say that narrow use
doesn’t fix meaning. I think there’s various things one can say here.
My view is that the central work of the activist project of conceptual
engineering can be done even by changing use widely enough. Usually doing that
will be enough to change the meaning. If it turns out that it doesn’t, changing use, I
think, is what we need for many practical purposes. I think this connects
to what Jared will be talking about this afternoon, on different notions of
meaning. I guess my inclination is, once you break down conceptual
engineering into design, implementation, and evaluation–concept design is
straightforward, concept evaluation is straightforward. They suffice for many theoretical purposes. Concept implementation requires changing
other uses, and is a difficult social project, but it’s possible. Okay, I think
my time is up. So some conclusions. Conceptual engineering includes, or at
least should include, heteronymous and the novo conceptual engineering, as well
as homonymous and conceptual re-engineering. In my view de novo
conceptual engineering is often the most fruitful, especially for theoretical
purposes. Homonymous conceptual engineering, I think, is also very
important, especially for practical purposes, and especially for social purposes that go well
beyond philosophy. I’m very strongly inclined think that conceptual
importance doesn’t exhaust philosophy; theses have priority over concepts, and
the importance of conceptual engineering really, to some considerable extent, derives
from its role in what we might call thesis engineering. So, thank you.
[Audience applauds] Vera Flocke: I think I’m gonna do what’s known as
the Anthony Gillis method, and assign numbers. And then you know when it’s
your turn. So if you have a question raise your hand. 1, 2, 3, 4, 5. 6, 7. Okay, everyone
should be able to ask their question. Man 1: Thanks, Dave. Very useful distinctions.
I mean I guess to introduce a new example, to me one of the most prominent examples of
de novo engineering is the concept genocide. That word indeed exists. That’s a good case where you can ask
what de novo means exactly, because the concept has a definition in terms of previously
available concepts. So it’s compositionally available in some sense. But in some sense
highlighted as new, as important for certain purposes. And this connects up with
what you were saying about theses. But the idea is, you know, Lemkin noticed that
there was a phenomenon that had not been picked out. It had certain features, he thought
those features were important for legal purposes, moral purposes, and so on. And so he
introduced the concept in order to name that. But it seems as though that–so that
general phenomenon, where you notice a phenomenon–of course there are many phenomena–
there are murders committed on a Tuesday, you could introduce a word for that–but there, I mean
although you might have introduced a new concept, it’s not clear
what use is the word. So it looks as though this is just, in some sense–
I mean science, right? I mean– that what is the distinctive thing has
to do specifically with concepts– Chalmers: Oh. I would have thought that conceptual
engineering is extremely important in science and in all kinds of social domains. So the question wouldn’t
even so much be what is so distinctive about concepts here, the question would
be what is so distinctive about philosophy here. And then I think the
answer might be nothing. I think you know I think conceptual engineering in
philosophy is to some considerable extent continuous with conceptual
engineering in science, conceptual engineering in social domains, and so on.
Nonetheless, all three of those things are extremely interesting for a
philosopher to analyze, and the philosophical analysis of conceptual
engineering, I think, is a very distinctive project for the philosopher.
Man 1: Yeah, I mean the emphasis in the case of genocide is, as it were, noticing there is a phenomenon.
Chalmers: Phenomenon engineering. Man 1: Phenomenon engineering. A phenomenon that
has–about which there are important generalizations, and then you need a way of talking about it.
Chalmers: So there’s a phenomenon–well, there’s a category, really, of a phenomenon. If you
just sort of pointed to, you know, the Holocaust and said, okay well there’s a
phenomenon. It’s putting it under a category, a general category,
which seems to be the distinctive contribution of a word like ‘genocide’. You
point to a few paradigm examples and say, look at all the things they have a
common; look at their common moral role. So I guess I think if that’s category
engineering, that would be the relevant thing. And you can see categories as
things which are out there in the world. Property. Is it property engineering?
Well again, I think properties–we didn’t create the property. What happens
in this case is we come to stand in a certain epistemological or
semantic relation to that property. One way to describe that is to come to
possess a concept of that property. So for that reason I think the word
‘conceptual engineering’ is still apt for this kind of phenomenon. Man 2: Just following up on this. As Paul said,
sometimes you design and engineer the concept by giving a definition. But the skeptical worry is that,
if you’ve got the definition, you know, you’ve already got the concept. So you’re not
really creating a new thing. What’s going on isn’t the creation of the concept. The concept’s
there already. You had it already. It’s the implementation problem. You’ve got a word, and
you want people to pay attention, you know, you’re directing attention at something
by people using the word. So that suggests that the issue is about conceptual activism
versus conceptual engineering. And this seems more on the activism side.
Chalmers: Yeah, except I guess I’m inclined to be not so sure that
everything you can define from your existing concepts is a concept you
have already. I think the concept of supervenience–you know,
there’s a random person, who happens to have the concepts of modality and the
sharing of properties, already have the concept of supervenience? Maybe they’re
in a position to acquire it. I think, you know, Moore and Hare and Davidson and Kim made very
serious conceptual epistemic progress by articulating that definition, and thereby puttin
people in a position to actually possess and deploy that concept. Many cases
don’t involve definitions. Man 2: Isn’t that the activism side? Getting people
to use the concept. I mean, you’re right– I shouldn’t have said everybody’s got the concept.
Chalmers: Okay, there’s still the design side. What goes on in
the design side–let’s say Hare was designing the concept of
supervenience. Let’s say he never actually got anyone to possess the
concept because of externalism. When anyone says supervenience, it means the old
meeting of supervenience. Can’t remember what the old meaning of supervenience
is. I just remember it’s used in Quine’s autobiography. Anybody remember that? “Necking eventually
supervened as necking always does.” [Audience laughs] Chalmers: It was the autobiography of
his teenage years. I’m sorry. [Audience laughs] Chalmers: In this meaning of supervenience it means
something like ‘eventuates’. Anyway. Just saying that the–even after Davidson, Kim,
Moore, and Cornell supervenience still means that thing, then still Davidson,
Kim, Hare, Moore made a contribution by pointing us–by designing that
concept, by pointing us to it, to point to all the roles that it can play, even if
nobody ever actually comes to possess that concept. I think that would still be
some kind of progress. I agree that after that is activism.
Flocke: No, I’m not going to take more fingers. Annette, number 2.
Woman 1: Great, thank you, Dave. Flocke: You can have a hand. So then you’re number 9.
Woman 1: So I think I see what you’re getting at with this distinction between de novo and
re-engineering. But then that immediately sort of raised for me a question about the object of evaluation. So when we’re thinking
about deficient representational devices. So in particular, especially thinking about
sort of the social context. It strikes me that maybe
we should be thinking about frameworks of concepts, as being the deficient thing. So I’m thinking about Miranda Fricker’s
discussion of the concept of sexual harassment, and how this didn’t exist, and
thinking, well, part of what was so bad about this wasn’t just that there wasn’t this
concept of sexual harassment, but also the concept that people had of women, and
men, and flirting and things like this–this whole cultural framework that’s going on. So
in introducing the concept of sexual harassment, it seems like that also required
some shifting in these other concepts, and so there, again, it seems like maybe the
object of evaluation should be more the framework. And then it’s like de novo re-engineering.
I mean I suppose you can still think of de novo engineering, but it would be a much
bigger, radical project. Chalmers: Yeah. I think you’re absolutely right
that a lot of the time it is gonna be these interconnected frameworks
that really get re-engineered. And then you’re right–so either you–
I mean, well especially if it’s going to have to be heteronymous, then you have
to introduce a whole lot of new words. One thing that does happen sometimes in
some cases is to introduce a common prefix to [unclear] all these things that’s
re-engineered. In teleological terms, we use teleo-this, or things have to be
re-engineered in moral terms, you use moral– So there are ways of getting at
the project heteronymously, say with something like a prefix, but in
those cases I think you’re absolutely right that there are particular
practical attractions to the homonymous strategy. It’s still–I mean, whether
it’s de novo or re-engineering, I guess the question is, do you think the
old framework is fundamentally defective and should be gotten rid of. In some
cases I think maybe that will be the case. Maybe in the case of–
the whole framework is fundamentally immoral or unjust. Certainly for many
purposes we should get rid of it. In other cases I could imagine–you know,
here’s a re-engineered version of this framework, which is useful, but the old
one is still useful for some other purposes, and then you might want to keep
both around. Woman 2: Thanks. Yeah, so I was wondering
about the claim that theses have priority over concepts. I certainly agree that
conceptual engineering is not the only thing that philosophy is
about. But I was thinking of the difference between projects you might think of as more
descriptive versus projects you might think of as more normative. So in the case of supervenience, you might
think that defending physicalism requires supervenience.
Versus cases where you’re saying something like, we need to rethink what
race and gender are, or what assault could involve, or like whether injustice could–
Katherine Jenkins is pushing a lot that there can be ontological injustice. Those seem kind of like–I mean,
you could build theses around these, obviously. Sort of like, we ought to use concepts like
this because that’ll promote a better world or something. But I was wondering if there
might be a difference between the sort of priority theses involve when you have a
project that’s more descriptive versus more normative. Chalmers: Yeah. So I was thinking about–I made this
distinction between theoretical and practical projects. Here I was thinking about the
theoretical side. But of course there are theoretical normative projects as well as
theoretical practical projects. So certainly I think, you know, the concept
of–re-engineering the concept of woman or the concept of race has got a very
clear–and doing it homonymously–has got a very clear social use. One which I
don’t think really derives from theses. I think that seems kind of weird. It derives
from roles. It derives from the roles that categories of race, womanhood, and so on
play in our society. They have major roles, and yeah, we think it would be
better in some cases maybe for those roles not to be played at all. That’s
something like Anthony’s view on race. Or in other cases
something different should be to playing those roles. So
that’s–so the practical uses I don’t think I’d want to argue for the
primacy of theses. Something more like the primacy of roles. For the
normative theoretical purposes–I’m just trying to think, what
would be an example that we should, use, say, the word woman to express the
concept in terms of oppression or maybe the concept in terms of
identification. That would be a normative–a meta-linguistic normative
theoretical thesis. I still think maybe that thesis–I think, you know, the
importance of the concept would derive from the truth of that thesis,
which would derive from the fact that it plays some important role also. I think maybe
there’s a way for you to hang on to the primacy of theses here,
but I gotta think it through. Flocke: Number four. That was you? Okay.
Man 3: I’m wondering what you would say to someone like Mark Wilson, whose view seems to be that the
whole history of scientific language is simply different kinds of conceptual engineering, and that [unclear] philosophy of scientific language at least is the best place within a traditional semantic theory.
Something more pragmatist [unclear]– Chalmers: Yeah, so I’ve read a lot of Mark Wilson’s
work, and it’s true that he thinks conceptual engineering is very important, but is it
true that he thinks that’s somehow– that’s all of science, or just that it
plays a very important role in science. Man 3: He doesn’t think that there’s much for philosophy
of language to do except describe the history of science. Chalmers: I think that’s a more general thesis
of Mark Wilson’s. Not necessarily tied to– in general he thinks that
the philosophy of language should be the history of scientific language. All of
the interests of language will be brought out by looking at what goes on
in actual scientific cases. And then a further thesis–an awful lot of what goes
on with scientific language is conceptual engineering. And then apply the first
thesis: the best way to understand conceptual engineering is to look at the
history of science. I certainly think that you can get an awful lot of insight
into conceptual engineering by looking at history of science–but the stronger
thesis, that somehow this is all we should be doing as philosophers, I guess I’m
inclined to reject both the first thesis, that, you know, there’s no more to philosophy of
language than looking at history of science–for some pretty obvious and familiar
reasons. But I’m also inclined to reject the second thesis. I think there’s a lot going on
with language in history of science, and conceptual engineering is one central
and crucial part of it, but not the only part. Flocke: Number five.
Man 4: I’m number five but my question’s been asked. Flocke: Okay, then number six.
Man 5: Thanks. Just to pick up on a question that was already asked, I guess I was–in a sense maybe the
internalist challenge to conceptual engineering. So one way of seeing all this is that we already have
the basic concepts, what we’re doing is just we’re putting them together in new ways to–well,
yeah, I guess the challenge then would be something like, well what we’re really doing
is we’re not conceptual engineering; what we’re doing
is we’re belief engineering. So what we’re doing is we are–so the hard part of conceptual
engineering is getting people to use the terms. This might just be construed as getting people to have
the beliefs we already have. And when you’re designing a concept, you might think of
that as, well what is guiding me when I’m designing a concept. Well I’m
looking at my beliefs, or looking at the world. I’m trying to figure out what
category of mine is the fruitful category for generalizations or whatnot. I’m not actually
sure this is a challenge, but anyways I just wanted to frame that.
Chalmers: So this is a version of the primacy of theses thesis, except for mental states, where we have
the primacy of beliefs thesis. It’s like, let’s engineer–if we’re thinking about engineering
mental entities, let’s engineer mental entities that can be
used in useful beliefs. Is this–are you thinking within philosophy or more generally?
Man 5: I guess–I guess the challenge would be to just like–
conceptual engineering as a category distinct from other kinds of belief-changing processes.
So you might distinguish between conceptual engineering on the one hand, you
know, like the arguing over theses– I’m not sure actually that this is correct but I’m curious whether–
yeah, whether you think that there is a clear line between, sort of, arguing with
theses and conceptual engineering. Chalmers: I don’t think there’s a clear line.
So I think–I mean one thing we’re trying to do in science and philosophy and many other
domains is to understand things and figure out the truth. Sometimes it involves recombining
old concepts, but sometimes it turns out to get real insight you need a new
concept. In the linguistic domain that might involve coming up with a new word;
in the mental domain that might involve coming up with a new concept. I guess I’m
inclined to think that–again the importance of the new concept is going
to be not just–hey, cool concept, but the work it can do for you
mentally. And some considerable part of that will involve beliefs. I mean maybe it
will also involve understanding, there may be some domains that might involve
connections to action. Maybe it’s going to lead to better or more just outcomes.
So again I’m inclined to think concepts are gonna be of instrumental utility, and one
very central part of that will be its role in the formulation of beliefs. And then–I
mean you’re right that nothing here is absolutely new. It’s like, we’re trying to
figure out the world, sometimes we come up with new categories and new concepts in the
in the service of beliefs. Flocke: Sigurd had a follow up.
Man 6: Yes, you sort of answered this now. I was also puzzled by the primacy of theses over concepts claim
you made. And I just wanted to sort of add to the discussion the idea that
there’s a–for any sort of argument about theses, there’s always a
standard of goodness or correctness having to do with what concepts you’ve
used to articulate those theses that doesn’t have directly to
do with the truth of the thesis that you’re arguing for. It’s
like present everywhere, and then it’s sort it very hard to–
Chalmers: Could you give an example of this kind of evaluation of concepts
that you have in mind? Man 6: Well, you could
criticize a statement for being vague, using a vague concept, and then in that case–I just
can’t see how you can separate out the concept from the thesis that it’s used in.
Chalmers: Yeah, let me try something here. I mean, mean why is a vagueness of the concept
used in a thesis a problem here? Because it makes the thesis vague. Just say I offered a
thesis which is really precise, and then I did ‘and P or not P’ on the end, for a
vague concept. I don’t think you could criticize that for using a vague concept, precisely because it
didn’t really matter, because the thesis itself wasn’t vague–let us assume a
semantics where the thesis itself comes out comes out non-vague. There
it seems to me it’s not such a bad problem. The problem of the–the fault of the
vagueness of–the reason why vagueness of concepts matters here is that it
brings about vague theses. The reasons why an inconsistent concept
might be a problem is they bring about inconsistent on non-existent
theses. Insofar as these things don’t permeate up to the level of theses, I’m
inclined to think those problems with concepts aren’t so bad. Now that’s probably
too strong a claim–I could probably be convinced–maybe there
are some concepts that it’s just immoral to think with, for example. I mean
I’m dubious about whether there is such a–these concepts like bosch, which have these moral
inferences. But if they did, that might be an example of a concept that it’s somehow
immoral or problematic to think with. Not because of any roles they play
in theses but because of bad things they do to you. Maybe vagueness and
imprecision could be–I mean some people think that, for example,
naturalness is an example here, that naturalness is just desirable, in
its own right, as something for a concept to express. But I’m inclined to think
naturalness is desirable because it’s productive and fruitful with respect
to theses. It helps you articulate laws and principles. If it wasn’t–if
we had something natural and something unnatural that were equally
productive, then it’s not clear to me that unnaturalness would be a problem.
Flocke: Laura also had a– Woman 2: Just a followup on this
point. One way in which you could have conceptual effects might be that it
triggers certain sort of stereotypes, or you tend to see–even though
you’ve got a determinate extension like with a slur term, you’ve gotten to see the
subject through the lens of a particular stereotype zone. It triggers
cognitive effects that are not really truth conditional effects.
Chalmers: Yeah, so that would be sort of a practical effect of certain concepts,
given–maybe tied to our nature as non-ideal reasoners. Maybe an ideal
reasoner could deploy those concepts without those effects. Maybe
you’re doubtful about that. I guess I’m inclined to think that the
question’s gonna be, are those effects constitutive of the concept or not?
If they’re not constitutive, then it’s not exactly a defect of the concept the way it’s
embedded in our own cognitive system. If it is constitutive then
maybe we’re getting closer to those concepts like bosch and so on. Flocke: I think we are at number seven now. Man 7: So I want to hear a little bit more
about how you think about the connection between the de novo case and [unclear]. So I was thinking,
well, what about that, and to describe roughly how I ended up thinking about it–I think the way
you set it up by focusing on this term ‘engineering’ is a little cheaty, because you know it’s
really just a label, it’s not very descriptive, and you can use other terms, so looking at
sort of compositional [unclear]– Chalmers: Externalism! The word–
[Audience laughs] The words are your masters,
not your servant. Man 7: So here’s a phenomenon, okay?
Let me just describe a phenomenon. So there are these philosophers who try to improve concepts like truth or belief or freedom. And that’s continuous with
things that–the concept of salad has changed over time.
In biology the concept of gene, in economics the concept of currency. So
these changes happen all the time. That’s really interesting: how does that happen, what are
the things that happen, what are the things that change, how do those changes happen ,to what
extent are they intentional, how is topic preserved through those changes. Now I said
a whole bunch–you know and the other question is, why do you keep the
same lexical item, when is it appropriate to change it. Now I’ve
described a bunch of really hard questions,
and I don’t think that if I answer those questions, I’ve answered the de novo
questions. So that’s why they’re different projects. I could do a really good
job answering those questions, and there would still be these unbelievably hard
questions about de novo. So they’re kind of different. Chalmers: Yeah, so I totally agree there
are distinctive questions about the two cases, and there are also interesting
questions about their union that they have in common.
I guess I’m interested in all three of those questions. You’re especially
interested in in the questions which are specifically about conceptual
re-engineering. That’s great. Somebody ought to be interested in–Carnap’s
project was conceptual re-engineering. Man 7: Do you think the answer to the re-engineering
issues I just raised would help you to solve– Chalmers: I’m almost certain that some of them would
help. That some of the questions you’re asking– Man 7: How do you–
Chalmers: Give me the list of ten questions. [Audience laughs] Man 7: How do those changes gradually happen in
all the examples I gave you, when is topic preserved, when is it appropriate to
preserve the lexical item, and when should it be changed. Let’s just start with those.
Chalmers: Okay, when is it appropriate to preserve a lexical item and when should it be changed–that’s
clearly a question which applies to both homonymous and heteronymous
conceptual engineering. Man 7: But the de novo case doesn’t have a
lexical item, so the issue doesn’t arise there. Chalmers: Yeah. I think
this is gonna overlap with the question of to what extent is the old concept useful,
and should it be kept around, and to what extent do we end up with plural concepts, and to what
extent should we end up with just a single concept and get rid of the other one.
And that is very much also a question about when is de novo conceptual–when is
pure conceptual re-engineering appropriate, which is getting rid of the old
one and replacing it with a new one, versus adding a new concept–when is that appropriate.
Anyway, I wouldn’t want to deny there are a lot of really interesting questions about conceptual re-engineering,
but I think if you go through, question by question, you’ll find (a) de novo conceptual engineering is a really
interesting phenomenon; (b) there are a lot of interesting and distinctive questions about that; and
(c) an awful lot of the questions about conceptual re-engineering and about
de novo conceptual engineering actually have a common answer. So it doesn’t make sense to study these two
things in complete isolation from each other. I guess from your perspective the
key thing would be the third one. Flocke: So we have three minutes left, and I would like to squeeze in two final questions by number eight and by number nine. Man 8: My question requires some
setup. I’ll try to be quick about it. I have a question about standards for
success in a re-engineering project. So let’s assume that
you and Andy were re-engineering the concept of belief in that 90s paper. And then we ask,
what good does that do? Clark’s typical sort of response is to say,
it opens up new pathways for research in cognitive science, and things like that.
But when one goes back and sort of looks at the work of
the folks doing the leading work in the 90s on situated cognition and
whatnot, it doesn’t look like they really had commitments here. Sometimes they talk about
[inaudible], sometimes they talk about it as part of the cognitive system. It doesn’t really much. So I’m
worried about the evaluation here, particularly with regard to the scientific case. Do we know enough about, like, how experimental
design is generated, or how people get ideas for a quantitative analysis of
data, to really make any claims about which of these ways of thinking about
belief is what drove the scientific progress. It seems like there might be a
sort of of disconnect here between– Chalmers: Yeah, it may be that
Andy and I have somewhat different attitudes about this. For Andy
the primary purpose was the use in cognitive science. For me that’s one purpose, but
among many. And I think what we’re finding in thinking about the extended
mind is having–it’s used for all kinds of people dealing with technology and
society, for reasons that go way beyond the science. People’s relationship to their
smartphones, or their internet search engines. And so I found that the
applicability of this thesis is just much–I mean Ned–it’s Ned who said that
back when we wrote the paper the thesis was false, but it’s since become true.
[Audience laughs] Chalmers: I don’t think that’s because of its role within science. To assess the role within
science–I don’t know, I see all this enormous literature these days on 4E
cognition, where extended cognition gets at least to be one of these. And many
people claim that it’s playing a central role in the science. I do agree that many of the points
in science where someone appeals to extended cognition could often be
reinterpreted as an appeal to embedded cognition. It’s kind of my own view
that any explanatory use of these things could probably–there’s gonna be a nearby
explanation where it’s done in non-extended terms. May just be less fruitful. Anyway. So I think to assess that we really have to engage in the Mark Wilson
project of looking at the science, but I do think there are uses outside science.
Man 9: Yeah, I wanted to talk about the challenge to de novo conceptual engineering that
both Pauls pushed, which was something like, a lot of these paradigms
seem to involve a case like rigid designator or supervenience, where you
introduce a new word in, but it’s a concept that’s really a concept for the
conjunction of some old concepts. And that looks a bit mysterious, so maybe everything
is on the practical side of getting people to use that word. I want to
suggest that there’s a different type of thing that might go on where you introduce
a new concept, and you don’t give any type of definition, and you kind of fail
at doing so. There might be lots of cases like this in the history of science. There’s
Newton and Leibniz were wondering about the notion of a limit, and kind of offer
these definitions but they aren’t very good.
Maybe there’s a concept that they have, but they can’t really articulate it. That seems like
there’s another type of de novo conceptual engineering, but it
doesn’t face this kind of challenge, and that’s got to be possible. I mean the other
alternative is very strong nativism that all
our conceptual resources are there from the start, and all we can do is
staple them together. Chalmers: So what’s an example of your wholly–your other kind, which isn’t–the
non-stapling kind of conceptual engineering? Man 9: It was supposed to be the
notion of a limit from Newton and Leibniz. I mean it’s much more mysterious how
that works. But it’s gotta be possible. Chalmers: Then there was a definition. I mean, there was
a concept of element and associated things like continuity. And then at a certain point we get a definition
of those things. And I guess it’s an interesting question whether it’s a new
concept. I think probably to some extent it’s at the
very least a precisification. Now the new concept, the epsilon-delta definition of
limits and continuity–well, I mean it’s not so clear to me that’s different in
kind from supervenience. It’s something that someone has stapled together from
things we understood. You know, the math of epsilons and deltas is is pretty well
understood. What’s new is (a) the stapling and (b) the connecting them to that role,
to that old, fuzzy notion we had of limits and continuity. I think that’s kind
of analogous to what goes on with, say, this precise concept of supervenience, and at
least attempting to connect that to the old, fuzzy concept of materialism or
physicalism. So there’s something which is engineered, which is
fairly precisely. What’s kind of fuzzy and vague is the role, and what’s
fruitful is somehow the connection between the two.
So maybe that’s not getting to the bottom of what you
think is going on in that. Flocke: Let’s thank our speaker!
[Audience applauds]

1 thought on “FCE Conference // David Chalmers, “What is Conceptual Engineering?”

Leave a Reply

Your email address will not be published. Required fields are marked *