E. Thompson

Evan Thompson
Reply to Commentaries
Let me express my deep thanks to the contributors for taking the time
to read my book, Mind in Life, and for writing their thoughtful com-
mentaries, from which I have learned a great deal. Special thanks are
due to Tobias Schlicht, whose hard work and dedication made this
volume possible. In what follows, I will respond singly to each con-
tributor (in alphabetical order) and do my best to address their main
points. My replies to the commentators will be longer or shorter
depending on the points they raised. (Unless otherwise noted, all par-
enthetical page references are to Mind in Life.)
Daniel C. Dennett
I would like to begin my response to Dennett on an autobiographical
note. In 1990–91 I spent a year as a postdoctoral research fellow at
Dennett’s Center for Cognitive Studies at Tufts University. At that
time I was in the last stages of writing The Embodied Mind (Varela et
al., 1991), and I was also at work on several papers and a book on col-
our vision (Thompson et al., 1992; Thompson, 1992; Thompson,
1995). Dennett was running a seminar based on the manuscript of his
book, Consciousness Explained. I learned a huge amount from
Dennett that year, not just from his seminar, but from his exceptionally
generous mentoring, which included numerous helpful conversations
and critical reactions to my writing, as well as many introductions to
leading philosophers and scientists. Since that time and to this day,
Dennett’s writings have stood for me as a model of how to do philoso-
phy in dialogue and collaboration with science, and his views on many
issues have served as a critical foil for my own thinking.
Correspondence:
Evan Thompson, Department of Philosophy, University of Toronto, Toronto, ON
M5R 2M8, Canada Email: evan.thompson@utoronto.ca
Journal of Consciousness Studies, 18, No. 5–6, 2011, pp. ??–??

2 E. THOMPSON
What Dennett and I share and where we differ can already be seen
by comparing The Embodied Mind and Consciousness Explained
(these two books were written around the same time, so we were not
able to take account of each other). Both books tackled the problem of
consciousness or lived experience from the vantage point of cognitive
science, and both discussed many of the same topics — the nature of
the self and personal identity, the temporality of conscious experi-
ence, the brain as a complex system, colour perception, and evolution,
among others. But whereas my co-authors and I advocated phenomen-
ology and first-person methods of examining experience as a necessary
complement to cognitive science, Dennett promoted heterophenomen-
ology; whereas we criticized adaptationism, Dennett defended it; and
whereas we advanced an embodied approach, Dennett upheld
functionalism.
In writing Mind in Life, I found myself again confronting these dif-
ferences and wanting to work through them in light of recent work.
Hence I gave special consideration to Dennett’s views throughout
Mind in Life, specifically when I discussed evolutionary theory, the
mental imagery debate, and the place of phenomenology in cognitive
science and the scientific investigation of consciousness. Although
much of what I wrote was critical, the criticisms reflect how important
and fruitful I find Dennett’s work for my own efforts.
Mind in Life was written in the wake of the death of my other men-
tor and close family friend, Francisco Varela, who had guided me
intellectually and personally since I was 15 years old. One way to
mourn a loved one’s passing and to try to keep their memory alive is to
take up their causes and fight for them. Reading Mind in Life now —
ten years after Varela’s death and four years since the book’s publica-
tion — I can see how that emotion and sense of purpose played a
strong role in shaping the form and content of some of what I wrote. I
refer specifically to my treatment of the issues that divided Varela and
Dennett — autopoiesis versus selfish genes, the autonomy perspec-
tive versus reverse engineering, and neurophenomenology versus
heterophenomenology. I suspect this driving emotion lies behind the
‘rathering’ that Dennett reads in some of what I wrote. Were I writing
Mind in Life today, I would try to give a more nuanced and balanced
presentation.
Nevertheless, when it comes to Dennett’s commentary, I have to
say that it seems to me to amount mostly to opinion and rhetoric rather
than argument (there’s a ‘rathering’ for you). Moreover, he seems to
misunderstand many of my points, and he sometimes states what he
REPLY TO COMMENTARIES 3
takes to be criticisms of my view whereas they are actually points I

made myself and took some time and care to discuss.
In what follows, I aim to identify the rhetoric, point out the misun-
derstandings, and re-state the substantive disagreements, so the reader
can judge for herself the relative merits of our positions.
Dennett’s rhetorical strategy is to identify himself with ‘ortho-
doxy’, while styling himself a ‘reform-minded critic’, and to brand me
as a ‘radical’ or ‘revolutionary’. He first used this strategy years ago in
his two book reviews of The Embodied Mind (Dennett, 1992; 1993).
Although these labels might have made some sense in the early 1990s
— and did admittedly fit my co-author Varela’s sense of himself as a
radical Chilean contesting the hegemony of Anglo orthodoxy — they
no longer seem apt. Cognitive science, philosophy of mind, and theo-
retical biology have changed considerably over the past decades, so
that the contemporary scene contains a plurality of theoretical per-
spectives instead of a monolithic ‘orthodoxy’ challenged by a few
‘radicals’ and ‘revolutionaries’. For this reason, at the beginning of
Mind in Life, I present the history of cognitive science from cognit-
ivism to connectionism to embodied dynamicism as a diversification
of perspectives and methods growing from the progressive recogni-
tion of the importance of context, lived experience, and temporal
dynamics for understanding cognition. As I write: ‘In contemporary
research, all three approaches [cognitivism, connectionism, embodied
dynamicism] co-exist, both separately and in various hybrid forms’ (p.
4). Nowhere in Mind in Life do I set myself up as a ‘revolutionary’ bat-
tling ‘orthodoxy’. These categories have little meaning for my pro-
ject, however convenient and comfortable Dennett may find them.
What does have meaning for my project is to give voice to certain sci-
entific and philosophical traditions that are invaluable for understand-
ing mind and life — traditions that have been neglected in certain
quarters and that provide important correctives against a number of
misguided tendencies in contemporary thought (e.g. genocentrism in
evolutionary theory, what I call ‘informational dualism’ in biology
and cognitive science, and the neglect of serious and careful phenom-
enology in the science of consciousness).
Dennett’s rhetorical strategy misdirects the reader away from seri-
ous consideration of the issues. The ‘revolutionaries’, he says, ‘are not
really so revolutionary after all’, and ‘reform-minded critics —
myself among them — have already pointed out the caveats that
pre-empt these assaults on orthodoxy’. Here we see the familiar ploy
of those who do not wish to confront seriously another way of think-
ing that challenges their own: acknowledge the rival viewpoint with
4 E. THOMPSON
an air of noblesse oblige, circumscribe its force by claiming to have

allowed for it already, congratulate yourself for being so undogmatic
and fair-minded, and then go back to doing things exactly as you had
been doing them before.
Let me respond now to the specific points Dennett makes about the
four themes in my book he chose to discuss.
1. Autopoiesis
First, a clarification. Dennett misreads me when he writes that I pro-
pose ‘autopoiesis as a radical new foundation for evolutionary the-
ory’. Autopoiesis is one key element in the account of life I offer;
others are evolutionary-developmental biology (‘evo-devo’), devel-
opmental systems theory, and theories of biological self-organization
(besides that of autopoiesis). I do not make autopoiesis foundational
for these other theories; instead, I weave them together into an
enactive perspective on evolution. At the beginning of Chapter Five, I
distinguish three complementary approaches to characterizing life —
the evolutionary, the ecological, and the individual (pp. 95–7). The
concept of autopoiesis targets life at the individual level and aims to
characterize the minimal organization necessary and sufficient for a
system to be living (there are problems with the sufficiency claim,
however, as I discuss on pp. 122–7). Later in the chapter (pp. 118–9), I
state that life at the individual level always has to be seen as ecologi-
cally embedded; and, at the beginning of Chapter Seven (pp. 166–7), I
make the point that the single, individual organism here and now is an
abstraction from both the organism as an ecologically embedded
life-cycle and from the organism as a member of a reproductive and
evolutionary lineage. I follow with a section devoted to explaining the
links between autopoiesis, reproduction, and heredity, in order to
show how the characterization of life at the individual level relates to
the characterization of life at the level of reproductive populations. In
the rest of the chapter, I use both the theory of autopoiesis and devel-
opmental systems theory to criticize genocentrism, and I use ideas
from theories of biological self-organization to criticize certain mis-
characterizations of natural selection. Dennett’s statement that I use
the theory of autopoiesis ‘as a radical new foundation for evolutionary
theory’ is thus an inaccurate simplification of what I write.
Dennett writes that the theory of autopoiesis is virtually synony-
mous with Tibor Ganti’s chemoton theory; that both basically amount
to a generalized and deepened version of the cell theory; that the the-
ory of autopoiesis does not contain anything that would particularly
inform a mainstream cell biologist; and that ‘it doesn’t predict any-
thing in biology that hadn’t already been well understood by earlier
theorists, or dissolve any puzzles that had been bedevilling those
theorists’.
Dennett is right that there are similarities between Ganti’s concept
of the chemoton and Maturana and Varela’s concept of autopoiesis;
we could also add Robert Rosen’s metabolism-repair (M,R) systems
to the list. (In Mind in Life, I discuss the relationship between
autopoiesis and M,R systems, but not the chemoton theory; Pier Luigi
Luisi, 2010, however, has reviewed the relationship between auto-
poiesis and the chemoton.) But Dennett’s dismissive assessment of
autopoiesis (and, by extension, perhaps also the chemoton theory)
misses the mark. The cell theory states that the cell is the fundamental
structural and functional unit of living organisms, and that all cells
arise from pre-existing cells. But this theory does not explain the orga-
nization that makes the cell an individual, how that self-producing
organization could be instantiated as a network of processes; and how
that kind of organization could in principle self-assemble from other
simpler processes and structures. All three approaches — autopoiesis,
M,R systems, and the chemoton — take this problem of the living
organization as their explanatory target. As I explain in Mind in Life,
Darwinian thinking, in its classical and modern molecular forms,
missed this problem about organization (a problem going back to
Kant’s discussion of the organism as a ‘natural purpose’). Here is a
case where, contrary to Dennett, the problem was indeed ‘unnoticed’,
not simply ‘underestimated’ (an assessment I justify in some detail in
Mind in Life). Today the problem occupies a central place in research
on the origins of life and synthetic biology (Luisi, 2010), as well as
artificial life and computational biology (see the 2004 special issue of
Artificial Life devoted to autopoiesis). In these research fields, the
theory of autopoiesis has underpinned work on the creation of
autocatalytic micelles as models of proto-cellular systems in the ori-
gins of life (Bachman et al., 1992), and it has inspired and guided a
significant body of work in artificial life (McMullin, 2004). More
generally, the theory of autopoiesis has helped to shape theories and
models, and make predictions, in research on the origins of life (Luisi,
2010), the chemical synthesis of minimal self-producing systems
(Bachman et al., 1992; Luisi, 2010), the computational simulation of
self-producing systems (Bourgine and Stewart, 2004; McMullin,
2004), and the modelling of autonomous agency in natural and artifi-
cial systems (Barandiaran et al., 2009). Mind in Life reviews much of
this work. Does Dennett really think that all this hard and original
6 E. THOMPSON
work was already well understood by earlier theorists and does not
help with any of the puzzles that bedevilled them?
Dennett writes that there are exceptions to my claim that ‘a cell
stands out of a molecular soup by creating the boundaries that set it
apart from what it is not’, and he presents the following as examples:
the boundary is semipermeable; which things count as inside and
which things count as outside is not always clear; there are transition
zones; when one autopoietic system enters another, it could be an
invader or a symbiotic ally. Here Dennett presents something as if it
were a criticism when it is actually a point I make myself. First, as I
point out, ‘boundary’ is equivocal. It can refer to the material-spatial
boundary of a membrane, or it can refer to the topological-functional
boundary that is determined by the system’s organization. As I write:
‘taking “boundary” to mean only a unicellular semipermeable mem-
brane or even a multicellular epidermal layer seems too restrictive
(plants and insects do not have a skin). Rather, the crucial matter is
that the system produce and regulate its own internal topology and
functional boundary, not the particular physical structure that realizes
this boundary’ (p. 107). I discuss this point in relation to both the issue
of whether multicellular organisms count as first-order autopoietic
systems (or instead as autonomous systems dependent on the
autopoiesis of their cellular constituents — see pp. 105–7) and the
issue of whether Gaia (the Earth’s ecosphere) counts as an autopoietic
system (as Lynn Margulis claims — see pp. 120–1). Here it is also
worth mentioning that I do not hold, contrary to Dennett’s suggestion,
that ‘only autopoietic systems can be the proper (literal, underived,
etc.) bearers of various biological predicates’. Social systems such as
ant colonies, beehives, and primate bands are not autopoietic; further-
more, it is an open question whether multicellular organisms or Gaia
qualify as genuine first-order autopoietic systems. Second, the exam-
ples Dennett gives are not exceptions to the point I make about the cel-
lular membrane; they point instead to the need to consider that
boundary as plastic and as a participant in what I later call (following
Ezequiel Di Paolo) the system’s ‘adaptivity’ (the system’s ability to
regulate itself in relation to its milieu — see p. 148). I develop this
point at greater length in more recent writings on the relationship
between the enactive approach and the extended mind theory in cog-
nitive science (Thompson and Stapleton, 2009; see also Di Paolo,
2009).
Another place where Dennett neglects what I write is when he asks
why we cannot see systems both as sources of their own activity, spec-
ifying their own domains of interactions, and as transducers or
functions for converting input instructions into output products. Of

course, we can look at systems in both ways, as I say myself: ‘What
counts as the system in any given case, and hence whether it is autono-
mous or heteronomous, is context-dependent and interest-relative.
For any system it is always possible to adopt a heteronomy or exter-
nal-control perspective, and this can be useful for many purposes’
(p. 50). Dennett also neglects the immediately following sentences:
Nevertheless, this stance does not illuminate — and indeed can obscure
— certain observable patterns of behavior, namely, patterns arising
from the system’s internal dynamics rather than external parameters.
An organism dynamically produces and maintains its own organization
as an invariant through change, and thereby also brings forth its own
domain of interaction… A heteronomy perspective does not provide an
adequate framework to investigate and understand this phenomenon; an
autonomy perspective is needed.
Dennett gives no argument to counter this point, but it is the crucial
point at issue.
Dennett objects to my statement, ‘natural selection is not an exter-
nal force but the differential propagation of developmental systems’
(p. 202). (Here he also misleadingly frames the issue as ‘the major
question of whether autopoiesis provides a genuine alternative to
standard neo-Darwinism’ — a claim I do not make for autopoiesis on
its own, as I have already explained.) Dennett wonders what the mis-
take about selection is and whether anybody has ever made it, but here
again he neglects what I write and the literature I cite (see pp. 206–9).
The idea that evolution is a ‘field of forces’ and that natural selection
acts as an outside force on the units of selection — the standard neo-
Darwinian view — was explained in detail years ago by Eliot Sober in
The Nature of Selection (Sober, 1984). My point is not that this con-
ception is mistaken but that it is limited because it neglects the self-
organizing dynamics of the objects of selection; when these dynamics
are taken into consideration, the objects of selection have to be seen
also as generators of selection; in this way, the idea of selection is no
longer adequately described as an external force (Weber and Depew,
1995; 1996, have explained this conceptual development in detail).
Furthermore, although I was not able to discuss this controversy in
Mind in Life, for the past few years one of the most interesting issues
in the philosophy of biology has been whether natural selection
should be interpreted statistically as a bias in the mathematical aggre-
gation of births, deaths, and matings — the view propounded by my
colleagues Mohan Matthen and Dennis Walsh — or whether it should
be interpreted as a cause of evolution. The enactive view of selection I
8 E. THOMPSON
present in Mind in Life goes well with the statistical interpretation of

natural selection, because it sees natural selection not as a cause but as
the differential retention of inherited variation (or in developmental
systems theory talk, as the differential propagation of developmental
systems), which itself has many causes. In any case, clearly these
issues about selection are important and substantive ones, but Dennett
engages none of them in his commentary; he just acts incredulous that
anyone could see things differently from the way he sees them.
Dennett objects to my saying that Dawkins’ concept of an evolu-
tionary arms race is a questionable metaphor; in Dennett’s view, this
concept is ‘one of the most predictively fruitful insights in evolution-
ary biology’. Here Dennett seems to miss my point. When I write, ‘the
notion of an evolutionary “arms race” is merely a questionable meta-
phor taken from the realm of human affairs and projected onto the
interactions between certain species’ and that ‘It is an entirely
observer-relative description’ (p. 205), what I mean is that ‘arms race’
is a loaded way of characterizing what can be more neutrally
described as positive feedback in certain coupled dynamical systems
in biological evolution (e.g. co-evolving genetic regulatory net-
works). What is predictively fruitful is not this metaphor but the math-
ematical models of the dynamical interactions. The metaphor is
questionable because it projects a complex, human sociopolitical phe-
nomenon onto biological evolution; in this way, it confuses the inter-
pretive framework we use to look at phenomena with the phenomena
themselves.
Dennett also neglects the context of my remark. That context is a
critical assessment of the adaptationist idea that living systems
become better ‘designed’ or ‘adapted’ to their niches through evolu-
tion by natural selection (e.g. through cumulative selection and evolu-
tionary ‘arms races’). Thus, immediately after my critical remark
about the ‘arms race’ metaphor, I make the point that evolutionary the-
ory provides no general variable property of ‘adaptedness’. Instead, it
uses a variety of significantly different technical measures of ‘fitness’.
Thus, talking about living systems becoming better ‘adapted’ to their
environments through evolutionary ‘arms races’ has no clear meaning.
Dennett says nothing in response to these points. Instead, he chal-
lenges me to show ‘a single instance in which autopoiesis (or develop-
mental systems theory) has predicted or explained biological effects
on a similar scale’. This challenge is misplaced for several reasons.
First, the concept of autopoiesis has helped to shape theories and mod-
els, and make predictions, in a variety of fields, as mentioned above.
Second, developmental systems theory is not a theory in the sense of a
specific model that makes predictions to be tested against other mod-

els; it is a general theoretical perspective, with roots in multiple exper-
imental traditions in developmental biology and developmental
psychology. Finally, to make a (valid) tu quoque point: the ‘arms race’
metaphor is not itself predictive or explanatory, for the reasons just
indicated.
Dennett finds unbelievable my charge that he confuses heuristics
and explanatory frameworks with the phenomena themselves. Yet, as
we have just seen, this is precisely what he does with the ‘arms race’
metaphor. In objecting to my charge, Dennett again seems to miss my
point and ignore the context of what I write. I am well aware that he
has ‘gone to considerable lengths over the years’ to show that ‘de-
sign-without-a-designer’ is not a contradiction, and that ‘the design
stance works exactly as well for organisms and their parts and behav-
iours as it does for artefacts’. But my assertion is not that the design
stance cannot be legitimately used for organisms or that the design
stance implies that organisms are the products of intentional design.
My accusation is that the design stance (reverse engineering) misses
something fundamental about living systems — their autonomy —
and that it is illegitimate to use this stance to argue that organisms
really are natural artefacts (heteronomous products of design without
a designer). The context of my remark is a consideration of the differ-
ences between functionalist and structuralist traditions in the history
of biology and their relation to the distinction between viewing living
systems as heteronomous (the reverse engineering approach) and
viewing them as autonomous. In the functionalist tradition, which
includes Paley’s natural theology, Darwin’s theory, and Dawkins and
Dennett today, the governing concept is that of design and the organ-
ism is likened to an artefact, either as the result of intentional design
(natural theology) or the blind watchmaker (Dawkins). Kant’s criti-
cism of applying the concepts of ‘design’ and ‘artefact’ to life, as well
as the criticisms made in the traditions of structuralism and Rational
Morphology, are not adequately answered by showing how there can
be design-without-a-designer (see pp. 129–40, 210–1). The crucial
point is that the concept of design does not bring into focus the auton-
omous organization proper to living beings. In Mind in Life, I build on
that point and argue (i) that reverse engineering is a heuristic or inter-
pretive stance that treats organisms as heteronomous systems; there-
fore, it does not provide the right kind of generalizations to talk about
biological autonomy; and (ii) to claim that organisms are natural
artefacts (the heteronomous products of design-without-a-designer)
because they can be interpreted from a reverse engineering perspective
10 E. THOMPSON
conflates that perspective with the phenomena themselves. Dennett’s

reminder that his design stance works for organisms as well as
artefacts just amounts to saying that organisms can be interpreted
from the design stance — a point I do not dispute (see, e.g. p. 460, note
22 for what I call ‘a nice case of reverse engineering’). What I do
argue, to repeat, is that interpreting them this way fails to reveal their
autonomy.
Dennett says I present ‘no support at all’ for the confusion I attrib-
ute to him (which is not the ‘elementary confusion’ he takes me to be
attributing to him, as I just explained). But the support is there in the
points I make in the surrounding context (as just detailed), to which
Dennett makes no reply; moreover, his remark in his commentary
about evolutionary ‘arms races’ displays the same confusion. In addi-
tion, I supply a supporting footnote (note 24, p. 461), in which I
explain (i) that Dennett’s term ‘Design Space’ already biases the dis-
cussion in favour of the functionalist or reverse engineering perspec-
tive, whereas the theoretically more appropriate and neutral term for
the phenomena under consideration is ‘Morphospace’, and (ii) that
Dennett’s view that modern biology would be impossible without
adaptationism mistakenly treats adaptationism as an a priori condi-
tion of biology instead of as a specific research programme. Dennett
does not reply to these points either.
Finally, when Dennett characterizes Maturana’s approach to auto-
poiesis as a reverse engineering approach, and says that my clarifica-
tion of autopoiesis reveals the ‘functionalistic rationales’ of that
theory, he neglects what I say on pages 144–5 about autopoiesis,
namely, that this theory is a theory of organization, not function, and
that, according to this theory, the notion of function has no explana-
tory value in characterizing a system’s autonomous organization.
2. Developmental Systems Theory

Dennett focuses in this section on my critical remarks about geno-
centrism and what I call ‘informational dualism’. The issues here con-
cern the role of genes in evolution and development, and the
conception of genes as carriers of information. Genocentrism or gene
selectionism holds that genes are units of digital information; that as
units of information genes have a special causal status in evolution
and development compared to all other non-informational factors;
that genes are ‘replicators’ that construct and control ‘interactors’
(phenotypes); and that genes are the primary units of selection in evo-
lution. Developmental Systems Theory argues that this conception of
genes is misguided: genes are not units of information in any sense of

‘information’ that would not also apply to non-genetic factors; the
replicator/interactor distinction is unhelpful (if a replicator is defined
as anything that reliably replicates, then there are many other
replicators besides genes; if a replicator is defined as an entity that
replicates itself through its own causal power, then the only replicator
is the reproducing organism or life cycle); and the unit of selection in
evolution is the life cycle.
It is striking that Dennett in his reply provides no arguments that
join this debate. Instead, he makes a series of complaints — that I
present a caricature of genocentrism (despite my having accurately
reported what he and Dawkins say); that I enlist the help of too many
theorists (despite the importance of their work, which he caricatures
as ‘revolutionary’ and offhandedly dismisses); and that I do not spell
out my claims about the misapplication of the concept of information
to genes (despite my detailed discussion of precisely this matter on
pages 54–7 and 179–87).
On pages 186–7 I quote two passages from Dawkins and one from
Dennett that express informational dualism. Dawkins’ passages con-
cern life and DNA; Dennett’s concerns consciousness and the identity
of the person through time. Each treats information as ontologically
distinct from its contingent material expression, as pre-existing that
expression, and as not affected by that expression. This way of think-
ing about information, I claim, reifies information and is structurally
isomorphic to vitalism and mind–body dualism.
Dennett says he does not understand my claim. Am I following
those ‘sciencephobes’ who use ‘the epithet “dualism” to attack any
science that uses the concept of information’? Clearly not; my book
contains a huge amount of science (and who are these unnamed
‘sciencephobes’ anyway?) Instead of attacking sciences that use the
concept of information, I use science to try to correct certain miscon-
ceptions about information (see pp. 57–8). Surely, Dennett writes, I
am not ‘claiming that the hardware/software dualism of computer sci-
ence is dualism of any objectionable (e.g. Cartesian!) kind’? Yes and
no. The hardware/software distinction of computer science does not
necessarily imply informational dualism, but the computationalist and
functionalist dualism of mind as informational software versus brain
as hardware is tantamount to informational dualism. As I write in
Mind in Life, ‘Genocentrism and computationalism… run on the same
conceptual fuel’ (p. 174; see also pp. 185–6). So, yes, I do object to the
hardware/software dualism of computationalism as well as to the mis-
application of this dualism to genes. Note that the computationalist
12 E. THOMPSON
version is exactly the dualism Dennett expresses in the passage I quote

on pages 186–7: ‘If what you are is the program that runs on your
brain’s computer… then you could in principle survive the death of
your body as intact as a program can survive the destruction of the
computer on which it was created and first run’ (Dennett, 1991, p.
430). Dennett seems blind to the fact that his dualism is indeed a dual-
ism of a Cartesian kind — a surprising oversight from someone who
coined the term ‘Cartesian materialism’ and who so trenchantly
showed how neuroscience and psychology can remain objectionably
Cartesian despite their being materialistic (ibid.). Of course, the
computationalist version is not a substance dualism, but it remains
markedly Cartesian in its disembodied conception of mental pro-
cesses (Descartes’ own view was actually more complex, as I discuss
in Chapter Eight of Mind in Life).
Dennett says I never spell out my claim that the hardware/software
dualism of computationalism is misapplied to genes. This charge is
unjust. On pages 180–2 I explain the claim and give two supporting
reasons with citations to the relevant literature (see also pp. 54–7).
First, hardware and software are independent of each other in a way
that DNA/RNA and the rest of the cell are not (the latter produce and
depend on each other autopoietically; the former do not); and second,
the notion of genes containing ‘information’ in the form of coded
instructions for developmental outcomes is faulty because it confuses
the causal specificity of the relations between DNA/RNA and amino
acids with coding for phenotypic design (see pp. 57 and 181–2).
Dennett makes no reply to these points. Instead, he offers what
seems to him ‘a very clear and unobjectionable way in which we can
draw the software/hardware distinction when discussing genes’. But
all he mentions are the familiar facts that DNA triplets specify amino
acids and that proteins are produced through the complex orchestra-
tion of RNA and ribosomes. For the reasons I explain in Mind in Life,
these molecular relations of causal specificity do not warrant a con-
ception of DNA sequences as software ‘coding for’ phenotypic char-
acteristics as hardware (see pp. 54–7, 180–2). Dennett, however, says
nothing about these reasons.
Dennett wonders whether anybody has ever subscribed to the myth
of the gene as a unit of pure information. He seems to have forgotten
Dawkins’s words:
After Watson and Crick, we know that genes themselves, within their
minute internal structure, are long strings of pure digital information.
(Dawkins, 1995, p. 17)
Genes are pure information — information that can be encoded,

recoded, and decoded, without any degradation or change of mean-
ing… We — and that means all living things — are survival machines
programmed to propagate the digital database that did the program-
ming. Darwinism is now seen to be the survival of the survivors at the
level of pure, digital code. (Dawkins, 1995, p. 19)
It is exactly this way of thinking that exemplifies the myth of the gene
as a unit of pure information.
Dennett complains that although I cite and criticize a passage in
which he discusses how genes do not carry information intrinsically, I
do not address his claim that it is nonetheless possible to make a prin-
cipled distinction between explicit (coded) and implicit (uncoded)
information. But Dennett does not answer the question that I raise
immediately after quoting him: ‘If information from the environment
is needed to make the genetic information informational in the first
place, then what is the ground for holding onto the genocentric tenet
that genes are the informational prime-movers?’ (p. 184). I then quote
a passage from Susan Oyama that details the problems with Dennett’s
conception of information; Dennett rejects the passage as ‘a series of
non-sequiturs’ but neglects to tell us what those non-sequiturs are.
In defence of the genocentric claim that genes (at least sometimes)
are informational prime-movers and cells their vehicles, Dennett
offers the familiar fact that tiny changes in genes can yield huge
downstream effects. But changes to non-genetic factors (such as
methylation patterns) can also have large downstream consequences
(see p. 177). Here too Dennett says nothing about the reasons and evi-
dence I present against the replicator/interactor conception of genes
and the cells that house them (see pp. 177–8, 197).
Dennett characterizes my marshalling of evidence and argument
from a number of important biologists and philosophers as ‘throwing
the kitchen sink at orthodox neo-Darwinism’, and he says that if this is
the best I can come up with, then neo-Darwinism must be in pretty
good shape. But two can play at that kind of rhetorical game: ‘If casu-
ally dismissing theorists, name-calling and caricature (“science-
phobes”, “radical biologists”, “a veritable Hall of Fame of would-be
revolutionaries of biology”), stating opinions as if they were argu-
ments, and not responding to evidence and reason are the best Dennett
can come up with, then my case must be pretty solid after all.’
3. Autonomous Meaning-Construction
The crucial issue here is what it takes for a system to qualify as having
genuine agency and a meaningful perspective on the world.
14 E. THOMPSON
I agree with Dennett that AI systems can in principle be designed to

‘muck about in the world and devise their own categories’. My claim,
however, or rather the fundamental hypothesis of the enactive
approach, is that such systems must be designed as autonomous
agents in the full-blooded and technical sense of ‘autonomy’ I specify
in Mind in Life (see pp. 44–6), and whose minimal paradigm is the
autopoietic cell. Since the writing of Mind in Life, this sense of auton-
omy has been defined more precisely with application to AI systems
by Froese and Ziemke (2009) and Barandiaran et al. (2009). As their
framework makes clear, and as I argue in Mind in Life (p. 160), repli-
cating molecules do not qualify as autonomous, and hence do not
qualify as having agency.
Dennett in his reply, however, claims that the autonomy perspective
‘is not required for the sense of sense-making’, and he reasserts his
claim that macromolecules are agents. Yet he provides no reasons to
back up this claim and to counter the reasons I give in Mind in Life for
rejecting it (see pp. 160–1). Instead, he gives two examples: ‘Think of
motor proteins — little porters trudging along on their actin or tubulin
highways carrying freight to where it is needed. Think of proof-read-
ing enzymes.’ Yet neither example counts as a case of agency. Agency
requires more than causally orchestrated behaviour — unless, that is,
we are willing to allow the word ‘agency’ to become so vague that the
planets going around the sun or sugar molecules crystallizing count as
agents. According to the enactive approach, agency requires at a mini-
mum that the system ‘manage the flow of matter and energy through it
so that it can, at the same time, regulate, modify, and control (i) inter-
nal self-constructive processes and (ii) processes of exchange with the
environment’ (Ruiz-Mirazo and Moreno, 2004, p. 240). Put another
way, agency requires at a minimum that the system meet the following
three conditions: (i) be an individual, in the sense of producing and
maintaining its own organization (‘individuality’); (ii) be the active
source of its interactions, in the sense of modulating the parameters of
its coupling with the environment on the basis of its internal
(self-organized) activity (‘interactional asymmetry’); and (iii) gener-
ate the norms for those interactions on the basis of its activity
(‘normativity’) (see Barandiaran et al., 2009). Replicating macro-
molecules such as DNA/RNA, motor proteins, and proof-reading
enzymes do not meet these three criteria, so they do not qualify as
agents in this rigorous and naturalistic sense. By contrast, chemotaxic
bacteria do qualify as agents. One of the basic errors of selfish gene
theory — which Dennett perpetuates — is that it metaphorically
projects a faulty conception of agency onto entities that do not meet

the minimal criteria for being agents.
4. Phenomenology and Heterophenomenology

The personal/subpersonal distinction and self-reports about experi-
ence. Dennett says he agrees with me that when one makes a
self-report about one’s experience, ‘One is describing one’s subjectiv-
ity at the personal level in a way that is completely noncommittal
about the subpersonal workings of one’s brain’ (p. 305). Nevertheless,
he also thinks that subjects make ‘unwitting reference’ to entities
whose status as real or fictional is to be decided by looking at what is
going on in their brains. Here we disagree. If the self-reports are prop-
erly descriptive in form (rather than conjectural about the underlying
causes of behaviour) — in other words, if they are proper phenom-
enological reports — then we should not interpret them as making this
kind of unwitting reference. For example, Dennett maintains that
heterophenomenology gives us ‘the leverage to discover that… since
there are no images being processed in the brain when subjects say
they are rotating mental images, their heterophenomenological
reports must be interpreted as unwitting fictions of a sort’. I maintain,
however, that when subjects say they are rotating mental images, they
are giving a personal-level description of their subjective experience
of visualizing how things look; they are not expressing beliefs or mak-
ing unwitting reference to what is going in their brain considered as a
cognitive system. Despite his claim to the contrary, Dennett’s way of
interpreting the subjects does amount to forgetting or abandoning the
personal/subpersonal distinction he introduced, for two reasons. First,
he evaluates purely descriptive self-reports about the phenomenal
character of experience on the basis of what is going on causally and
subpersonally in the brain; and second, he evaluates the reports by
looking to see whether the reported phenomenal content matches or
fails to match the representational format of the neural processes. To
put the point another way, my claim is that even if there are no picto-
rial representations in the brain, it does not follow that the intentional
content of imagery experience is fictional in Dennett’s sense. As far as
I can see, Dennett does not counter the argument I present for this
claim on pages 304–5; he just re-states the view my argument
challenges.
Intentional acts and intentional objects. Here Dennett reads ‘inten-
tional’ as meaning deliberate, instead of being directed toward an object.
According to phenomenology, conscious experience is intentional in the
16 E. THOMPSON
sense that it is about or directed toward an object, and every intentional

object implies a correlative mental act that intends that object. The object
as perceived implies a certain act of perception that perceives it that
way, the object as remembered implies a certain act of remembering
that remembers it that way, the object as imagined implies a certain of
act of imagination that imagines it that way, and so on.
My point in Mind in Life (pp. 305–6) was twofold. First, we need to
attend to both correlative sides of an intentional experience when we
analyse that experience phenomenologically. In the case of mental
imagery, for example, we need to characterize the intentional object
— the shape as visualized — and we need to characterize the inten-
tional act — the activity of visualizing (see pp. 291–7 for this analysis,
and pp. 297–303 for its application to the mental imagery debate).
Second, intentional acts such as perceiving, imagining, and remem-
bering should not be identified with believing, which is another type
of intentional act. Believing is what happens when subjects have to
make a judgment. Thus, in the case of subjective experience, a belief
about experience is what gets expressed when subjects are asked to
make and report a judgment about their experience.
First-person methods. I do not deny that heterophenomenology can
acknowledge the unavoidable need to make use of first-personal
modes of access to mental phenomena. As I write (p. 306): ‘There
seems to be nothing in the heterophenomenological method that disal-
lows using the first-person perspective in this direct phenomen-
ological way… if the material on which heterophenomenology goes
to work is first-person reports about experience, and if the production
of such reports sometimes requires that subjects attend to and describe
their experience, then heterophenomenology already depends on the
first-person mode of access to mental phenomena being put to work in
an experimental setting.’ My critical point is that heterophen-
omenology has had virtually nothing to say about the idea — a key-
stone of phenomenology — that conscious experience needs to be
explored from within the first-person perspective (either descrip-
tively, as in phenomenological psychology, or transcendentally, as in
transcendental phenomenology). Moreover, it has had nothing to say
about the proposal to use first-person methods of training attention
and awareness in order to sensitize individuals to their experience in
ways enabling them to describe it more precisely (compared to indi-
viduals without that kind of mental training). This proposal is a key
feature of neurophenomenology, in which one of the working hypoth-
eses is that reports from individuals with such mental skills are useful
for identifying the variability in the brain response on a trial by trial
basis, so that more precise correlations can be established between the

spontaneous flow of consciousness and spatiotemporal patterns of
intrinsic brain activity (Lutz and Thompson, 2003; Cosmelli et al.,
2007; Lutz et al., 2008a). I like Dennett’s liberal conception of
heterophenomenology as giving ‘the subject the best possible way to
let it all hang out’, but this allowance seems like lip-service given
heterophenomenology’s lack of consideration to first-person methods
(except to deride their use in ‘lone-wolf phenomenology’, which no
proper phenomenologist has ever advocated or practised). Hetero-
phenomenology on its own does not provide the first-person phenom-
enological procedures needed for a proper examination of subjective
experience; it presupposes them. For this reason, as I say in Mind in
Life (p. 307), heterophenomenology by itself must be deemed meth-
odologically incomplete.
With regard to the Lutz et al. (2002) study, which I prefer to think of
as a pilot study and not a ‘flagship attempt’, I agree that it has the limi-
tations that Piccinini (2010) mentions. Meanwhile, since the Lutz et
al. pilot study, other neurophenomenological studies using first-per-
son and second-person methods for probing experience have begun to
appear (e.g. Christoff et al., 2009; Farb et al., 2007; Lutz et al., 2008b;
Petitmengin et al., 2007), so this kind of research, although fledgling,
looks promising.
Experience versus beliefs about experience. Here the disagreement
is easy to state. My view is that properly descriptive statements about
experience should be interpreted as direct expressions or verbaliza-
tions of those experiences and not as beliefs about those experiences.
Interestingly, Piccinini (2010) proposes precisely this way of inter-
preting first-person reports as one of a number of changes for improv-
ing heterophenomenology. In his words:
What remains unclear [in Dennett’s procedures for doing heterophen-
omenology] is why we should assume that reports always express
beliefs about the target mental states rather than expressing the target
mental states themselves… A more prudent treatment of reports is at
hand: establish to the extent possible that a report expresses the target
mental state (e.g., an emotion, desire, memory, or what have you,
including a belief if that’s the target), then interpret the report as provid-
ing information about the target mental state. After all, that is what we
are trying to investigate. (Piccinini, 2010)
I would add only that one point of first-person phenomenological
methods is to enable subjects to express maximally and accurately
their target mental states, instead of having them make judgments and
thereby express their beliefs about their target mental states.
18 E. THOMPSON
For some examples of first-person/second-person methods being

used to gain access to implicit aspects of emotion and cognition, see
Farb et al. (2007); Nielsen and Kaszniak (2006); and Petitmengin et
al. (2007).
What’s in a name (phenomenology or heterophenomenology)?
Dennett says my own proposed method ‘is really heterophen-
omenology after all rather than a radical alternative’, and that if I will
adopt his method, then he will adopt my name for it — second-person
heterophenomenology. Let’s set aside the epithet ‘radical’ as a distrac-
tion from the main issue. If Dennett’s method now (i) permits using
first-person methods for examining experience; (ii) encourages their
use in certain contexts (e.g. for examining attention and emotion regu-
lation in relation to neural plasticity, or for relating spontaneous fluc-
tuations in subjective experience on a fine time-scale to neural
measures on a trial by trial basis); (iii) treats phenomenological self-
reports as verbalizations of target mental states rather than expres-
sions of belief about those mental states (unless there is good reason to
interpret them otherwise); and (iv) does not evaluate the truth or fal-
sity of the content of such reports by examining whether they match or
fail to match the representational format of the subpersonal neural
processes, then there is no longer any substantive difference between
the methods we propose. As for the name, ‘phenomenology’ seems to
me less cumbersome and misleading than ‘heterophenomenology’,
for as I write in Mind in Life (p. 307): ‘Phenomenology from its start
has already encompassed heterophenomenology (or its possibility).’
To close my response to Dennett, I think I have shown that Mind in
Life is hardly about setting up strawmen or poking certain authors
with a parody of their views. As for getting Dennett to dance the
tango, I wish I could make it happen, because I think we all can benefit
from learning some new moves.
Lucia Foglia and Rick Grush

In their commentary, Foglia and Grush ably defend an emulation
account of mental imagery against a simulation account, and they ask
whether I see the emulation account as a friendly clarification of my
enactive account or as a challenge to it. The question arises because
enactive accounts criticize representationalist theories of perception,
and the emulation theory appeals to representations, understood as
mental models. The short answer to their question is that I welcome
their account and see it as a friendly supplement to my remarks about
sensorimotor processes in mental imagery (see pp. 295–6, 298–9).

But let me say a little bit more.
First, the concept of representation is used in numerous ways in
cognitive science and the philosophy of mind, as Hutto helpfully
details in his commentary. I argue against representationalist theories
that separate perception and action, instead of recognizing their con-
stitutive interdependence, and that neglect the ways autonomous
agents bring forth or enact meaning in perception and action (see pp.
10, 58–9). Since the emulation theory does not require these typical
features of representationalism, my objections to representationalism
need not apply to the emulation theory.
Second, appealing to emulation models to explain mental imagery
still leaves open exactly how those models are realized in the brain.
With regard to this issue, I have doubts about the way Grush (2004)
applies control theory to the brain. Specifically, I doubt that the engi-
neering distinction between ‘controller’ and ‘plant’ (controlled sys-
tem) applies to the brain considered as a complex dynamical system
(see Kelso and Kay, 1987; Kelso, 1995; and Appendix B to Mind in
Life). This issue, however, goes well beyond Foglia and Grush’s com-
mentary, so I will not pursue it here.
Finally, Foglia and Grush’s argument that mental imagery requires
an internal model is reminiscent of Sartre’s (2004) idea that imagery
requires an ‘analagon’ that functions to direct the image conscious-
ness to its intentional object. In perceiving a picture, the analagon is
overt; in visualizing an object it belongs covertly to the image con-
sciousness (though not as an object inside consciousness). This inter-
esting correspondence between phenomenology and the emulation
theory, regarding the role of ‘models’ in imagination, would be worth
pursuing further.
Dan Hutto
Hutto aims to open up a dialogue between analytical philosophy of
mind and the enactive approach. His strategy is to show how teleo-
semantic theories of content need to be modified in a variety of ways
that end up bringing these theories closely in line with the basic orien-
tation of the enactive approach. I welcome this dialogue, greatly
appreciate Hutto’s bridge-building efforts, and find myself largely in
agreement with his commentary.
My one caveat concerns something Hutto himself mentions — the
interpretation of evolution. Teleosemantic theories have traditionally
worked with a strongly adaptationist view of evolution, in which
20 E. THOMPSON
natural selection is conceived as an outside force acting on agents con-

ceived in heteronomous terms (see my response to Dennett). In con-
trast, the enactive approach calls attention to autonomous systems as
generators of selection, and to the need to understand selection as a
kind of dynamical stabilization resulting from many causes (see also
Oyama’s commentary). Hence, Hutto’s version of a modified teleo-
semantics — ‘teleosemiotics’ as he calls it — would also need to move
away from adaptationist views of evolution in order to find common
ground with the enactive approach. I see no reason why this move-
ment cannot happen, though I suspect the resulting teleosemiotic the-
ories would look rather far removed from their teleosemantic
ancestors.
Albert Newen
Newen thinks my ‘aims are much too ambitious’ and he criticizes the
path I chart from autopoiesis and sense-making to cognition and
consciousness.
1. Autopoiesis, Adaptivity, and Cognition

One of the core ideas of Mind in Life is what I call the ‘deep continuity
of life and mind’. There are different ways to express this idea. One
way is to say that to be a living system is also to be a sense-making
system, and thus that life is sufficient for mind. More precisely stated,
being an adaptive autopoietic system is necessary and sufficient for
being a living system; and being an adaptive autopoietic system also
suffices for being a sense-making system. To be a sense-making sys-
tem is to be a cognitive system, in a wide or broad sense of the term
‘cognitive’.
Newen challenges this thesis; in his view, adaptivity is the crucial
property for cognition, not autopoiesis. He seems to misunderstand
these concepts, however, including their relation to each other.
As the models of autocatalytic micelles and autopoietic tessellation
automata show, minimal autopoiesis (a semipermeable boundary
autocatalytically constructed by reactions remaining within the
boundary but also occurring at the boundary) is sufficient for individ-
uality, but not for behaviour in the sense of the system’s modulating
the boundary conditions and parameters of its coupling with the envi-
ronment on the basis of its internal activity. In more concrete biologi-
cal terms, minimal autopoiesis is necessary but not sufficient for such
basic phenomena of life as metabolic assimilation and accomodation,
or chemotaxis. Such phenomena also require what Ezequiel Di Paolo
(2005) calls ‘adaptivity’, namely, being able to modulate the auto-

poietic process in relation to conditions registered as improving or
deteriorating, viable or unviable, for the system. As it happens, every
naturally occurring autopoietic system we know is adaptive in this
sense; nevertheless, as the above models show, it is possible to con-
struct minimally autopoietic systems that are not adaptive.
The crucial point Newen misses, however, is that, in the case of an
adaptive autopoietic system, adaptivity is not an external add-on to
autopoiesis but a complexification of autopoiesis. In other words,
autopoiesis is not a mere background condition; autopoiesis is consti-
tutive of adaptivity. To state the point more generally, in the case of an
adaptive autonomous system (the paradigm of which is an adaptive
autopoietic system), autonomy is constitutive of adaptivity. For this
reason, it is not the case that the explanations for the sense-making or
cognitive abilities of living systems rely only on the property of being
an adaptive system and not also on the property of being an autono-
mous system.
The first section of Newen’s commentary contains a few other
problems:
(1) Newen’s declaration that bacteria, amoebae, and plants are not
cognitive systems is question begging, for I maintain that these organ-
isms are sense-making systems, and thus cognitive in a broad (but
well motivated) sense of the term.
(2) My reference to ‘internal self-production of the minimal sort’ on
page 129 is a reference to minimal autopoiesis; contrary to what
Newen seems to imply when he quotes me, living cells are not merely
minimally autopoietic but also adaptively autopoietic. Furthermore,
bacteria, the simplest naturally occurring autopoietic systems, are
highly adaptive, so it is not the case that such cells can survive only as
long as the environment is very stable.
(3) It is misleading to say that I accept Bourgine and Stewart’s
(2004) tessellation automaton ‘as a case of an autopoietic system that
is neither a biological nor a cognitive system’. The tessellation autom-
aton is a mathematical model of an autopoietic system, not an actual
autopoietic system.
2. Does Autopoiesis Plus Adaptivity Entail Sense-Making?

Newen argues against my claim that autopoiesis plus adaptivity suffices
for sense-making or what I call cognition in its minimal biological form.
For Newen, cognition requires that the system be representational, spe-
cifically that it operates on the basis of internal representations of goal
22 E. THOMPSON
states. Without such internal representations, he maintains, we have no

norms but only mechanisms.
First, let me point out an apparent misreading. I do not call sense-
making a ‘narrow sense of cognition’. Rather, I say that living is a pro-
cess of sense-making, of bringing forth significance and value, and I
describe sense-making as cognition in a broad or wide sense of the
term (p. 159).
Second, Newen’s view has the following problems: (1) he helps
himself to the notion of representation, as if it were obvious what this
notion means, and without addressing my reasons for rejecting it, at
least as this notion has been standardly understood (see Hutto’s com-
mentary for further discussion). From my point of view, his appeal to
representations as the mark of the cognitive is question begging; (2)
he presupposes that there is a meaningful way of contrasting mecha-
nisms and norms. Yet one of the main points of my discussion of
autonomous systems is that such systems are organized in such a way
as to be normative (see also Barandiaran et al., 2009). So the dichot-
omy between mechanistic reactions and norms does not apply to these
kinds of systems in the way Newen assumes; (3) such normativity is
not something inside the system (e.g. as the content of a representation
of a goal); it is a relational property of the system and its environment.
In sum, Newen’s criticisms in this section depend on assuming or
presupposing concepts and distinctions I do not accept and that I sub-
ject to criticism in Mind in Life. Newen does not address those
criticisms.
3. The Deep Continuity of Mind and Life

Newen thinks my deep continuity claim can be read in two ways, one
weak and trivial, the other strong and false. Neither reading gets at the
heart of the deep continuity thesis.
According to the weak reading, the organizational properties of
mind are an enriched version of those fundamental to life. Newen does
not make clear why this claim is trivial, but he does say that it is nei-
ther original nor controversial. I make no claim to originality. As for
being controversial, if ‘organizational properties’ refers (as it does
here) to self-organization, collective dynamics, circular causal pro-
cesses, autopoiesis, and so on, then the claim is certainly controversial
from the standpoint of cognitivist cognitive science as well as most
versions of functionalism (see Clark, 2001, p. 118, quoted on pp.
128–9).
According to the strong reading, ‘the principles of mind are essen-

tially constituted by the principles of life’. I do not make this claim in
Mind in Life. What I claim is (i) that the organizational principles of
mind are an enriched version of those fundamental to life, and (ii) that
there is an existential-phenomenological continuity of mind and life
(p. 129). This second claim is the one I develop through a reading of
Kant and Jonas.
But what about the question of whether the principles of mind are
essentially constituted by the principles of life? Newen seems to think
this claim is obviously false for two reasons. First, cognitive abilities
require representations of goal states. I have already indicated that I
reject this assumption (for reasons indicated in Mind in Life). Second,
autopoietic principles are mere ‘background conditions’ for cognition
but not essential to cognition. On this point, however, much more cau-
tious consideration is needed (see also the commentaries by Protevi,
Wheeler, and my replies).
I maintain that autopoiesis plus adaptivity is (i) necessary and suffi-
cient for life, and (ii) sufficient for mind. It follows that mind is neces-
sary for life. The question remains: is life necessary for mind?
On the one hand, I maintain that mind — sense-making, cognition,
and consciousness — requires that the system be autonomous (in the
precise sense of autonomy detailed in Mind in Life). I also allow that
there can be autonomous systems that are not themselves autopoietic
systems. So far, however, every autonomous system we know
depends constitutively on autopoietic constituents.
But what about AI systems and robots? AI and robotics have not yet
managed to create autonomous agents in the relevant sense of ‘auton-
omy’ (see Froese and Ziemke, 2009; Barandiaran et al., 2009). Hence
I disagree with Newen when he writes, ‘Modern robots with complex
behavioural abilities, e.g. in spatial navigation, have knowledge about
their environment but they are so far not autopoietic systems’. Because
these robots are not autonomous agents, they do not embody knowl-
edge about their environments; the ‘knowledge’ is attributed and in
the eye of the beholder (see Froese and Ziemke, 2009).
The crucial remaining question is whether autopoiesis is a neces-
sary constitutive feature of autonomy; in other words, can there be
autonomous systems that do not depend constitutively on auto-
poiesis? Put another way, does autonomy require metabolic self-
construction?
Although some might think the answer to this question is obviously
no, reasons can be given for thinking the answer might rather be yes.
For example, perhaps sense-making requires the kind of selfhood and
24 E. THOMPSON
concerned perspective on the world that comes from having con-

stantly to renew oneself metabolically in precarious thermodynamic
conditions. If there were this requirement, then autopoiesis would be
necessary for autonomy, and so life would be necessary for mind. I
postpone further discussion of this issue for my reply to Wheeler.
Were autopoiesis to turn out to be necessary for autonomy and
hence for cognition, it would not follow that appealing to autopoiesis
or to being a living system would suffice for explaining cognition.
Nothing I write in Mind in Life implies that I maintain the implausible
thesis that the principles of being a living a system suffice for explain-
ing all of cognition, so it is not clear to me why Newen seems to think I
believe this thesis.
4. Artificial Life
Here I need to correct two misunderstandings.
First, as already indicated, I do not maintain that the tessellation
automaton is an autopoietic system that is not a biological system; I
maintain that it is a mathematical model of minimal autopoiesis. I also
call attention to the question of whether autopoiesis can be completely
modelled this way when I discuss the relationship between auto-
poiesis and Robert Rosen’s metabolism-repair systems (p. 144).
Second, it does not make sense to say that ‘autopoietic organization
is a structure’. An organization is a set of relations among processes; a
structure is a concrete instantiation of those relations. The autopoietic
organization can be concretely realized in a variety of structures, and
any autopoietic system is constantly changing its structure (through
the metabolic turnover of its constituents).
5. Can Life Be Known Only by Life?

Jonas’s statement ‘life can be known only by life’ is a transcendental
statement in the following sense: it is about the conditions for the pos-
sibility of knowing life, given that we do actually have biological
knowledge (p. 164). Newen’s claim that I do not distinguish clearly
enough between knowledge in general and empathic knowledge
misses the point. My claim is that in order to recognize or bring into
focus a certain form or pattern of phenomena — the pattern of a
self-producing unity relating flexibly and adaptively to its environ-
ment — we need to embody that pattern ourselves, i.e. to be living
beings. I will not repeat Jonas’s argument (or my version of his argu-
ment) for this claim here. Instead, I will indicate why I think Newen’s
considerations against transcendental thinking have no merit.
First, the transcendental framework I am working with in Mind in

Life is not strictly Kantian; it belongs to transcendental phenomenol-
ogy following Husserl and Merleau-Ponty. I explain this conception
of the transcendental in Chapter Two. Newen’s three considerations
against transcendental philosophy, however, target Kant and do not
speak to transcendental phenomenology: (i) transcendental phenom-
enology rejects the postulation of an unknowable thing-in-itself; (ii)
transcendental phenomenology does not separate the transcendental
ego and the empirical ego, but maintains that the transcendental ego is
the empirical ego considered as a condition of possibility for inten-
tional consciousness (see Sokolowski, 2000); (iii) transcendental phe-
nomenology does not preclude an understanding of mental causation,
but views mental causation in terms of the motivational relations
between embodied intentional experiences.
Second, Newen’s statement that ‘modern philosophy of mind in its
majority’ has left behind transcendental thinking is simply false and
hugely distorts philosophy of mind. Transcendental considerations
abound in philosophers such as Wittgenstein, Davidson, Putnam,
Evans, and McDowell, to name just a few. Furthermore, one of the
most interesting and fertile philosophical developments in recent
years combines transcendental and existential phenomenology with
analytical philosophy of mind and philosophical psychology in order
to make headway in understanding phenomenal consciousness (e.g.
Zahavi, 2005).
6. Consciousness
Newen misreads my discussion of phenomenal consciousness and
attributes to me claims I do not make. I never claim that autopoiesis is
‘essential’ for phenomenal consciousness. Instead, I show how certain
existential structures of embodied experience have their roots in basic
structures of biological life.
Newen writes that although human feelings are ‘connected’ with
self-regulation, ‘there is no evidence that this is essentially so’. It is
not clear to me what he means by ‘essentially’. Does he mean that it is
conceptually or metaphysically possible for human emotions to occur
in the absence of how the human body affectively regulates itself in
relation to environmental events? What would be the argument for
this highly implausible claim?
Similarly, Newen claims that ‘sensorimotor couplings’ are not
essentially connected with phenomenal consciousness. Here I suggest
26 E. THOMPSON
he read Siewert’s commentary, which ably elaborates my view that

bodily movement is constitutive of perceptual consciousness.
Newen states ‘humans are already born with basic phenomenal
consciousness that is active before intersubjective interaction can
develop’. This statement overlooks the different aspects of inter-
subjectivity I discussed in Chapter Thirteen. The statement is true
only if ‘intersubjective interaction’ refers to social cognition involv-
ing perspective-taking and mutual self/other understanding (pp.
395–401); it is false if ‘intersubjective interaction’ includes the ‘pri-
mary intersubjectivity’ of affective resonance and sensorimotor cou-
pling (pp. 393–5; see also Gallagher and Meltzoff, 1996).
Newen also misreads my criticism of the zombie argument. First, I
do not grant the logical possibility of zombies. It is not obvious to me
at all that zombies are logically possible; to determine whether they
are would require that we fill in all the important details in the zombie
scenario, something proponents of the zombie argument never do. I
strongly doubt, for the reasons given in Mind in Life, that the zombie
scenario is conceivable (though I do not claim to have shown that it is
inconceivable). Second, I do not criticize the zombie argument in
order to show ‘the non-reducibility of consciousness’. I criticize the
zombie argument in order to reveal the highly problematic conception
of phenomenal consciousness it presupposes.
Newen also misunderstands my claim that ‘Consciousness, consid-
ered as epistemic base, is equivalent to the experiential acts by
which… objects are disclosed to us’ (p. 239). This is a transcendental
claim about consciousness as the condition of possibility for appear-
ance, including what appears to scientific observation; it is not an
empirical claim that ‘consciousness is the essential feature for unify-
ing the information into object representations as part of a meaningful
world’.
Nevertheless, Newen’s considerations do not undermine this
empirical claim, for several reasons. First, contrary to Newen, it is far
from obvious that ‘representations with [exactly] the same content
can be processed either consciously or unconsciously’. Second, in the
context of the studies involving pathology (visual agnosia, hemi-
neglect, blindsight), ‘conscious’ could mean accessible to verbal
report, or phenomenal but not accessible to report and action guid-
ance. Newen’s conclusion that these studies show that consciousness
is not required for certain types of processing is too quick unless it
carefully distinguishes between different senses of ‘consciousness’
and different ways that consciousness can be assessed and measured.
Finally, in every case of ‘unconscious’ processing he mentions, the
subjects are conscious in the sense of being awake and having a uni-
fied (though disrupted) phenomenally conscious field of awareness
(see pp. 351–2). So these ‘unconscious’ processes do not constitute a
case of perceptual processing in the complete absence of consciousness.
7. Enaction and Consciousness

In this section Newen is concerned with certain aspects of Alva Noë’s
view of perception. Since I do not make use of those aspects in Mind
in Life, I will not comment on them here, except to say that, to my
mind, Noë (2004; 2005) has already dealt satisfactorily with the
objection Newen raises. Instead, I urge Newen to read Siewert’s com-
mentary, which does an excellent job of showing how perceptual
experience is constituted through bodily movement.
8. Conclusion
Newen thinks functionalism and identity-theory can account for all
the phenomena discussed in my book, a preposterous claim consider-
ing the failure of these theories to provide any kind of comprehensive
and satisfactory account of the mind, let alone the specific phenomena
I discuss.
Susan Oyama
Oyama raises a number of important concerns about the relationship
between the theory of autopoiesis and Developmental Systems
Theory. I like her conception of these theories as ‘neighbours’ and I
agree that my claim of ‘complementarity’ for the two theories still
requires further working out in relation to the questions she raises
concerning internalism and causal asymmetry (the determination of
what counts as ‘inside’ versus ‘outside’, and the attributing of asym-
metrical causal roles to internal versus external factors). I cannot fully
address those concerns here, so I offer the following general remarks
as a way to keep the conversation going.
In Mind in Life, I write that autopoiesis (in a broad sense that includes
adaptivity) is the ‘self-production of an inside that also specifies an out-
side to which it is normatively related’, and thus that autopoiesis is best
seen as the ‘dynamic co-emergence of interiority and exteriority’ (p.
79). Yet I also immediately go on to say that ‘there seems to be an asym-
metry here, for it is the internal self-production process that controls or
regulates the system’s interaction with the outside environment’ (ibid.)
To support this point, I quote two philosophers and theoretical biolo-
gists, Alvaro Moreno and Xabier Barandiaran (2004), who write
28 E. THOMPSON
about what they call, following Varela, the ‘basic autonomy’ of life:
‘the (self) generation of an inside is ontologically prior to the dichot-
omy in-out. It is the inside that generates the asymmetry and it is in
relation to this inside that an outside can be established. Although the
interactive processes [and] relations are necessary for the mainte-
nance of the system, they presuppose it (the system) since it is the
internal organization of the system that controls the interactive rela-
tions’ (Moreno and Barandiaran, 2004, p. 17).
A number of authors, including Oyama in her commentary, have
expressed worries about this assertion of asymmetry between interior
and exterior. Donn Welton (2011) suspects it of being a kind of ‘bio-
idealism’, and argues that it unduly downplays the way the environ-
ment leads the organism into certain rhythms, behaviours, and internal
transformations — an environmental role he calls ‘affective entrain-
ment’. John Protevi (2010) wonders whether Varela’s notion of an
autonomous system ‘overemphasizes the individual as self-conserv-
ing product as opposed to individuation as always ongoing process’.
I am sympathetic to these helpful and friendly (or neighbourly) crit-
icisms, for a certain tendency to privilege interiority in autopoietic
discourse has always troubled me. I felt that worry in writing those
words in Mind in Life about the reciprocal yet asymmetrical relation
between interiority and exteriority, but I did not adequately address
the worry because of another argument I was trying to advance, spe-
cifically that the genuine interiority of life is a precursor to the interi-
ority of consciousness, and hence that the conception of nature
presupposed in standard formulations of the hard problem or explana-
tory gap for consciousness — namely, that living nature has no genu-
ine interiority — is misguided.
Here is another way to come at the issue about interiority Oyama
raises. On the one hand, I claim that the adaptive-autopoietic process
‘brings forth’ or ‘enacts’ what counts as the living being’s world, and
not the reverse; on the other hand, I claim that the living being and its
environment are ‘structurally coupled’, and interiority and exteriority
are ‘dynamically co-emergent’. So how do we resolve this issue of
asymmetry in the reciprocal coupling of living beings and their
worlds?
At this point I would like to inject an autobiographical remark to
indicate how long this tension has preoccupied me. Varela and I began
working together on The Embodied Mind in the late 1980s when I was
a graduate student. It was during those years that Varela introduced
into his work the terminology of organisms ‘enacting’ and ‘bringing
forth’ their worlds, rather than representing them (though this idea
was already implicit in his earlier work on autopoiesis with

Maturana). This way of talking worried me — precisely for its not
fully worked-out suggestion of some kind of idealism or construct-
ivism. So whenever Varela would write that the organism enacts its
world, I would try to rewrite the sentence to say that a world is brought
forth or enacted by the structural coupling of the organism and its
environment. My aim was to shift the emphasis away from the organ-
ism as the enactor of its world to the relational process of enactment.
Varela was happy with these changes, as they fitted better our other
sympathies with the Indian Buddhist concept of dependent co-origi-
nation (pratÌtyasamutpªda), also central to The Embodied Mind. Nev-
ertheless, my re-wording clearly did not deal adequately with the
tension, for the question of the asymmetrical versus symmetrical sta-
tus of the organism — or of the adaptive autopoietic process — in the
relational process of enactment remained unanswered.
Welton (2011) proposes a way to resolve this tension with his
notion of affective entrainment. He writes, referring specifically to the
requirement of adaptivity for sense-making:
Adaptation is much more than a dynamic adjustment allowing the
organism to get along better with its habitat according to internal
self-generated norms. It is also a transformation of the organism’s inter-
nal processes and norms according to the demands of an environment
that introduces ‘sense-producing’ or ‘sense-demanding’ requirements
of its own… [T]he environment that the organism opens or enacts is
also the world that entrains it and reflexively transforms both the pro-
cesses and the structure of the cell ‘reacting’ to it. (Welton, 2011)
In the case of bacteria, for example, the presence of sucrose exerts an
‘extrinsic’ control over chemotaxis; in dynamical systems language,
sucrose acts as an external control parameter, entraining the cells to
swim up-gradient. It is precisely this entrainment, Welton maintains,
that accounts for the status of sucrose as attractant.
I welcome and agree wholeheartedly with these points. Living as
sense-making is systemically generated and, as Oyama notes, system-
ically controlled. Living beings enact environments that pull them
along into certain rhythms, behaviours, and internal transformations
(this point becomes especially important when we remember that the
environment is always an environment of other living beings — bacte-
ria, for example, do not live in isolation but in microbial communi-
ties). In Welton’s words: ‘The organism enacts an environment as the
environment entrains the organism. Both are necessary and neither,
by itself, is sufficient for the process of sense-making’ (ibid.)
30 E. THOMPSON
But now comes the tricky point. What we have just said implies that
the relation between organism and environment is reciprocal, for each
acts as a control parameter for the other. But this kind of reciprocity
does not imply that their relation is not also asymmetrical, in the rele-
vant sense of asymmetry. Although the physical and energetic coupling
between a living being and the physico-chemical environment is sym-
metrical, with each partner exerting more influence on the other at dif-
ferent times, the living being typically modulates the parameters of this
coupling in a way the environment typically does not (Barandiaran et
al., 2009). Living beings, precisely because they are autonomous and
adaptive, can ‘surf’ environmental events and modulate them to their
own ends, like a bird gliding on the wind. ‘Interactional asymmetry’ is
precisely this capacity to modulate the coupling with the environment.
If we lose sight of this interactional asymmetry, then we lose the abil-
ity to account for the directedness proper to living beings in their
sense-making, and hence we lose the resources we need to connect
sense-making to intentionality.
It is crucially important, however, to realize that ‘boundary’ in this
context cannot be identified with any given spatial boundary, such as a
membrane, but refers instead to the system’s topological boundaries
as an autonomous network of processes. The way these processes are
structurally realized is plastic, both compositionally (what materially
composes them over time) and spatially (where they are located in
relation to spatially specified boundaries). The processes constitutive
of an autonomous network can incorporate external material
resources and extend beyond the biological membrane of the body, as
happens, for example, when a blind person uses a cane to perceive the
environment (see Di Paolo, 2009; Thompson and Stapleton, 2009).
Thus Gregory Mengel (as reported by Oyama) gets me right when he
says that my ‘internal-external distinctions are less about spatial
boundaries… than about selfhood, organizational closure, and the
context-dependence of causes’. Furthermore, as far as I can see, inter-
actional asymmetry in the above sense does not involve the kind of
arbitrary causal privileging that Developmental Systems Theory
criticizes, because such asymmetry does not accord a special causal
status to processes just because they happen to occur on one or the
other side of some spatial boundary.
John Protevi
I find the links Protevi makes between my project in Mind in Life and
Deleuze’s writings fascinating, but I do not know Deleuze well enough
to respond, so I will focus on what he calls the ‘question of panpsych-

ism’ (see also Wheeler, who raises the question of panpsychism).
Protevi thinks that although my conception of the deep continuity
of life and mind escapes from the Cartesian problem of the relation
between the mental and the physical, it raises the problem of the emer-
gence of life and mind from non-life. He wonders whether I am too
restrictive in my conception of mind when I trace mind back to living
as sense-making. Moreover, given that I work with the concepts of
processes and networks as webs of processes, what is to stop me from
embracing the kind of process panpsychism we find in Whitehead or
Deleuze?
To address this issue I want to compare Jonas and Merleau-Ponty,
because it is precisely on this issue about matter and life — or what
Merleau-Ponty (in The Structure of Behavior) calls the physical order
and the vital order — that I follow Merleau-Ponty and not Jonas.
Jonas contrasts the wave and the organism. The wave he takes to be
a material aggregate, which, as ‘an integrated event-structure’, has no
ontologically emergent status. He writes that to the wave ‘no special
reality is accorded that is not contained in, and deducible from, the
conjoint reality of the participating, more elementary events’. In other
words, Jonas accepts analytical and ontological reductionism for
physical phenomena. What he then argues is that this kind of
reductionism fails in the case of the organism, which is ontologically
emergent. Life, as he puts it, is thus an ‘ontological surprise’.
Now, if we follow this line of thought, then I think we do face a seri-
ous life–matter problem, analogous to the mind–body problem. How
does life emerge from non-life? The panpsychist argues that we can-
not make good on this invocation of emergence, that it is ultimately
mysterious. Hence the options would seem to be either some kind of
dualism or some kind of panpsychism.
But this line of thought is not at all the one we find in Merleau-
Ponty. Already in The Structure of Behavior, Merleau-Ponty rejects
analytical reductionism for physical forms like waves, soap bubbles,
and convection rolls (see pp. 72–3). As he writes, ‘The genesis of the
whole by composition of the parts is fictitious. It arbitrarily breaks the
chain of reciprocal determinations’ (Merleau-Ponty, 1963, p. 50).
Consider also this passage, which I quote in Mind in Life (p. 72):
[E]ach local change in a [physical] form will be translated by a redistri-
bution of forces which assures us of the constancy of their relation; it is
this internal circulation which is the system as a physical reality. And it
is no more composed of parts which can be distinguished in it than a
melody (always transposable) is made of the particular notes which are
32 E. THOMPSON
its momentary expression. Possessing internal unity inscribed in a seg-

ment of space and resisting deformation from external influences by its
circular causality, the physical form is an individual. It can happen that,
submitted to external forces which increase and decrease in a continu-
ous manner, the system, beyond a certain threshold, redistributes its
own forces in a qualitatively different order which is nevertheless only
another expression of its immanent law. Thus, with form, a principle of
discontinuity is introduced and the conditions for a development by
leaps or crises, for an event or for a history, are given. (Merleau-Ponty,
1963, p. 137)
As I state in Mind in Life, this description of physical form as introduc-
ing a principle of discontinuity and the conditions for development by
‘crises’ has been borne out by René Thom’s ‘catastrophe theory’, which
mathematically describes abrupt transitions and qualitative discontinu-
ities in physical systems, and by Jean Petitot’s extension of Thom’s
work to a morphodynamical ‘physics of phenomenality’, which aims to
bridge the gap between the microphysical substrate and macro-
physical forms.
What we find in Merleau-Ponty is a reconceptualization of matter,
life, and mind, one that does not bring mind down into the domain of
microphysical processes nor equate mind with information transfer
and self-organization, but rather tries to show how the notion of form
as dynamic pattern or individuation process can both integrate or
bridge the orders of matter, life, and mind, while also accounting for
the originality of each order. This is the path I try to follow in Mind in
Life and not panpsychism.
Nevertheless, I admit that my characterization in Mind in Life of life
as autopoiesis plus cognition can be read as equating mind and life,
and hence as opening a door to the panpsychist line of thought. What I
would now rather say — and these remarks pertain also to issues
raised by Newen and by Wheeler — is that living is sense-making and
that cognition is a kind of sense-making. A wave or a soap bubble is an
individuating process but not a sense-making one, because it does not
modulate its coupling with the environment in relation to virtual con-
ditions and norms. A unicellular organism is a self-individuating and
sense-making being but not a cognitive one if by ‘cognition’ we mean
being intentionally directed toward objects as unities-in-manifolds
having internal and external horizons (call this the phenomenological
sense of cognition). Ultimately, what matters to me is not to fix the
meanings of the words or concepts ‘matter’, ‘life’, ‘mind’, ‘cognition’,
and so on. Instead, my aim is to see whether we can chart multiple
passages back and forth between those orders that we conceptualize —

in different ways and at different times — as matter, life, and mind.
Charles Siewert
Siewert focuses on my treatment of the explanatory gap. He raises a
challenge for my treatment and develops an account of embodied con-
sciousness in order to meet the challenge. I welcome this account and
think it takes an important forward step that builds on what I write in
Chapter Eight of Mind in Life.
Siewert finds a lacuna in my argument that a proper phenomen-
ological account of perceptual experience shows that the same per-
ceptual functions that occur in our world could not occur in a world in
which there were no experience. I argue that it is on the basis of our
kinaesthetic experience of our own body that we are able to perceive
objects in space as unities in and through perspectivally varying
appearances, and thus that bodily self-experience is constitutive of the
perceptual function of individuating objects in space. This perceptual
function, I argue, would not occur in a world in which there were no
bodily self-experience; hence there could not be a zombie that was
functionally equivalent to us in its perceptual abilities. Siewert won-
ders, however, whether I have not stipulatively built a tie to bodily
self-experience into the concept of perceptual functioning. My claim,
however, is that our perceptual abilities to individuate and track
objects in space depend constitutively on bodily self-experience; it is
thus a claim about the functional role consciousness plays in percep-
tion. In making that claim, I do not, as far as I can see, build bodily
self-experience into the concept of the relevant perceptual functions.
In any case, the lacuna Siewert sees is the need for a proper demon-
stration that perceptual functioning ‘cannot be factored into a “phe-
nomenal, experiential bit” and a “bodily movement bit” with no more
than a contingent causal link between them’. I agree that I have not
provided such a demonstration, and I am grateful to Siewert for work-
ing to provide it.
Siewert presents an analysis of perceptual experience that shows
that looking and touching are indissolubly both consciousness and
movement, and hence that our bodily engagement with the world can-
not be factored into a phenomenal component and a movement com-
ponent that are only causally and contingently related. I greatly
admire Siewert’s analysis. Had I had thought of it myself I certainly
would have included it in order to bolster my argument that perceptual
experience is so tied to embodied activity that the same perceptual
34 E. THOMPSON
functions occurring in our world could not occur in a world in which

there were no experience. Siewert’s defence of the ‘Embodied Con-
sciousness View’ (Section 3) and his application of this view to the
explanatory gap are very helpful additions to my efforts.
I also admire the way Siewert locates the embodied-consciousness
view in relation to other views of the explanatory gap. The only place
where we seem to differ concerns zombies. Siewert distinguishes
between a zombie world functionally equivalent to ours and a zombie
world physically equivalent to ours. My ‘phenomenological critique’
of zombies targets the functional version but does not show that a
physical zombie world is inconceivable. Nevertheless, it is far from
clear that a physical zombie world is conceivable. To believe that such
a world is conceivable we need to know how to imagine this world in
such a way that arbitrary details about it can be filled in without any
incoherence arising. To my knowledge, no proponent of zombie
conceivability ever provides this kind of guidance. One of my aims in
Chapter Eight of Mind in Life is to show how unsatisfactory this way
of thinking is, especially when it is tied — as it typically is — to
phenomenologically impoverished ways of thinking about conscious-
ness in relation to embodiment (see pp. 233–4).
Siewert also raises a question at the end of his commentary. The
question concerns how the ‘inwardness’ of biological life relates to
the inwardness of subjectivity. He wonders whether inwardness in the
sense of maintaining a boundary between inside and outside is neces-
sary for inwardness in the sense of being a suitable focus of empathic
regard, and he asks whether I mean to show how the first can be built
up to the second, and whether we need to close that gap in order to
explain consciousness.
My remarks here will be brief; for related considerations see my
replies to Newen, Protevi, and Wheeler. My view is that life or living
being already involves a kind of inwardness that goes beyond what
can be captured in an external conception of material structure and
mechanical function. That kind of inwardness is not mere spatial
inwardness (as marked by a membrane) but rather the inwardness of
immanent purposiveness and being normatively related to the envi-
ronment. I maintain that this kind of inwardness is a precursor to sub-
jectivity, but I do not show how the one can be built up to the other;
that task requires much more work and belongs to what I call the
‘body–body problem’. A full account of consciousness and its relation
to biological life would require an explanatory bridge from the one
kind of inwardness to the other, but there is plenty of work that can be
done on consciousness without that kind of bridge being in place, as
Siewert’s own presentation of the embodied-consciousness view

amply demonstrates.
Robert Van Gulick

Like Dennett, Van Gulick sees my rhetoric in Mind in Life as ‘radical’
or ‘iconoclast’ in relation to the ‘mainstream literature’, which I ‘at-
tack’. For the reasons indicated in my response to Dennett, I do not
share this way of looking at my efforts in relation to the complex ter-
rain of philosophy of mind, cognitive science, and theoretical biology.
Although my views are certainly iconoclastic from the perspectives of
functionalism, cognitivism, and adaptationism, I prefer to think of
myself as a pluralist who draws from the many approaches that popu-
late contemporary science and philosophy.
Van Gulick offers teleofunctionalism and biosemantic theories of
meaning as approaches that share many of my ideas about the deep
continuity of mind and life. At a general level, I agree, but this resem-
blance strikes me as superficial. There are two crucial differences
between these theories and my approach (see also Hutto). First, these
theories treat organisms as heteronomous systems, not as autonomous
ones; and second, these theories assume an adaptationist account of
evolution instead of the enactive one I propose.
Van Gulick thinks it odd that I do not mention non-reductive
physicalism in my list of mainstream options for dealing with the
explanatory gap. But non-reductive physicalism is arguably an unsta-
ble position that ultimately results in either eliminativism or property
dualism. So it does not provide a genuine alternative to either dualism
or materialism. As for my position in relation to other approaches to
the explanatory gap, see Siewert for a very useful assessment of the
lay of the land.
Van Gulick cites Lycan’s teleofunctionalist approach as convergent
with mine. In the Lycan passage Van Gulick quotes, however, Lycan
proposes a reductionistic account that would explicate the mental in
non-mental terms through a strategy of homuncular decomposition.
My approach is very different. First, I do not believe that mental terms
can be reductively translated into non-mental terms. Second, I main-
tain that homuncular decomposition fails for complex systems (see
Appendix B). Third, Lycan’s conception of function is extrinsic func-
tion, not the immanent purposiveness of autonomous organization
(see pp. 144–6).
There is also a much larger background philosophical difference
between these theories and my approach. Non-reductive physicalism,
36 E. THOMPSON
biosemantics, and teleofunctionalism advance a kind of naturalism

that denies the transcendental status of consciousness and subjectiv-
ity. The kind of naturalism I propose, however, recognizes the tran-
scendental status of consciousness and reconceptualizes nature from a
transcendental standpoint.
Van Gulick states that few if any ontological physicalists subscribe
to mereological reductionism because they assert that ‘the properties
of wholes are fully determined by the intrinsic properties of their basic
parts plus their mode of composition and arrangement’. Like non-
reductive physicalism, however, this position is arguably unstable,
and leads either to mereological reductionism or to property dualism.
My account of emergence is meant to offer a third way between these
problematic alternatives (see Appendix B).
Michael Wheeler1
Wheeler raises a number of important questions about the ‘deep conti-
nuity thesis of life and mind’ through a probing analysis of the rela-
tions among the concepts of autopoiesis, autonomy, adaptivity,
sense-making, cognition, and teleology. I am grateful for this analysis
because it gives me the opportunity to formulate the interconnections
among these concepts more clearly than I did in Mind in Life.
Let me start with the following schematic presentation. I maintain
the following theses:
(1) Autopoiesis and adaptivity are individually necessary and jointly
sufficient for life. In other words, life is autopoiesis plus adaptivity.
(2) Autonomy and adaptivity are individually necessary and jointly
sufficient for immanent purposiveness (each part being both a
product and producer of the other parts, so that the whole system
is a self-organizing whole) and sense-making (behaviour or con-
duct in relation to significance, valence, and norms that the sys-
tem itself brings forth or enacts on the basis of its autonomy).
(3) Sense-making is cognition in a wide sense of the term; or, to put
the point another way, sense-making is the basic mark of the
cognitive.
(4) Autopoiesis is the paradigm case of autonomy, in the sense that it
is the best understood case and the minimal case of an autono-
mous organization (more on ‘minimal’ below).
(5) Thus, autopoiesis and adaptivity are jointly sufficient for imma-
nent purposiveness and sense-making.
[1] I wish to thank Ezequiel Di Paolo for helpful discussion of Wheeler’s commentary.
(6) Thus, life is autopoiesis plus cognition (any living system is an

adaptive autopoietic system, hence an adaptive autonomous sys-
tem, hence a sense-making system, hence a cognitive system).
(7) Thus, cognition is necessary for life.
We thus arrive at one sense of the deep continuity thesis of life and
mind: wherever there is life there is mind.
The crucial remaining question is whether autopoiesis is necessary
for autonomy and hence also necessary for sense-making. If the
answer to this question were yes, then autopoiesis would be necessary
for cognition, and we would arrive at another sense of the deep conti-
nuity thesis, namely, that life is necessary for mind, or wherever there
is mind there is life.
To answer this question we need to be clear about what autopoiesis
requires. Wheeler states that ‘autopoiesis is autonomy plus material-
ity’, by which he means that ‘to be autopoietic, an autonomous system
must, through its own endogenous self-organizing dynamics, produce
and maintain a material (or physical) boundary which distinguishes
that system as a material (or physical) unity in the space in which it
exists’. Although this description is true, it does not specify the crucial
definitive feature of the autopoietic organization that makes an auto-
poietic system a particular kind of autonomous system. What makes
the system autopoietic is not its self-produced material boundedness
as such, but rather that the relations constituting the system are rela-
tions between processes of molecular transformation, including those
that make up the boundary. To put the point another way, the physical
boundary of a cell should not be confused with its organizational
boundary (see also my replies to Dennett, Newen, and Oyama). A
molecule traversing the cell membrane is spatially inside the bound-
ary that contains most of the processes that make up the autopoietic
system, but this location does not by itself determine whether the mol-
ecule is part of the autopoietic system. The molecule can become part
of the autopoietic system only if it can become integrated into the net-
work of processes that sustain the relation of closure among them-
selves (e.g. if the molecule becomes a participant in metabolic
reactions). If the molecule does not participate in such processes or if
it disrupts them, then the molecule remains ‘foreign’ (non-self).
When we ask whether autopoiesis is necessary for autonomy and
sense-making, it is crucial to realize that what we are asking is
whether the autopoietic organization is necessary for autonomy and
sense-making, not whether the ‘dual materiality of the systemic
38 E. THOMPSON
boundary and the systemic mode of existence’ is necessary for auton-

omy and sense-making.
In addition, we need to refine what ‘necessary’ means. I maintain
that there are autonomous systems that are not themselves first-order
autopoietic systems, such as immune networks, the nervous system,
insect colonies, animal societies, and primate bands. Multicellular
organisms are a tricky case (see pp. 105–7); so too is Gaia (the Earth’s
ecosphere). In any case, in one important sense, I maintain that auto-
poiesis is not necessary for autonomy (viz. it is not necessary that an
autonomous system itself be autopoietic). Notice, however, that in all
these cases, the autonomous system depends constitutively on having
autopoietic components.
Wheeler is thus doubly misguided, I think, when he writes ‘such
materiality [belonging to autopoiesis] is apparently expendable by the
time that biology gets as far as the nervous system’. On the one hand,
as we have seen, it is not the materiality per se of an autopoietic sys-
tem that is the relevant feature, but rather the way the system realizes
the autopoietic organization in the molecular domain. In the case of
the nervous system, the system depends constitutively on components
(nerve cells and glial cells) that realize the autopoietic organization in
the molecular domain. On the other hand, it is far from obvious and
arguably false that the materiality of the nervous system is expendable
(think of how the electrical signalling properties of neurons are chem-
ically realized, and how the nervous system depends crucially for its
functioning on complex molecular cross-talk with the immune and
endocrine systems).
Wheeler might respond that the materiality of the nervous system is
expendable in the sense that the cognitive functions the nervous sys-
tem implements could be realized in a materially different structure.
This thesis of ‘multiple realizability’, however, should be treated as an
empirical hypothesis that could well be false (Shapiro, 2004). Here we
need to distinguish between the following two ideas — multiple
realizability and compositional plasticity. If a functional property
(e.g. being a watch) can be implemented in different physical mecha-
nisms with different causal properties (analogue watch, digital watch,
sundial), then that property is multiply realizable. Mere difference in
physical composition, however, is not sufficient for multiple realiz-
ability; the compositional difference must entail a difference at the
level of mechanisms and causal properties. Suppose we constructed
an artificial nervous system. If the artificial ‘neurons’ realize func-
tional states by virtue of the same electrical properties as biological
neurons, then the artificial brain does not realize the functional states
in a relevantly different way (Shapiro, 2004). We can thus distinguish

between the following two empirical hypotheses: (i) the multiple
realizability hypothesis: the cognitive functions the nervous system
implements can be realized in different physical systems with differ-
ent causal properties; (ii) the embodied hypothesis: the cognitive func-
tions the nervous system implements can be realized only in systems
having the causal properties of the biological nervous system.
According to the first hypothesis, the materiality of the nervous sys-
tem is expendable; according to the second it is not. Deciding between
these two hypotheses requires evaluating the empirical evidence — it
cannot be decided on the basis of conceptual considerations alone.
We thus arrive at another sense in which we can ask whether auto-
poiesis is necessary for autonomy, namely, whether autopoiesis is a
necessary ingredient for autonomy, or to put it another way, whether
there could be an autonomous system that did not realize its autonomy
through autopoietic constituents and thus did not depend constitu-
tively on autopoiesis.
Before addressing this question, I need to clear up a few points aris-
ing from Wheeler’s reading of my views.
First, when I speak of ‘minimal autonomy’ or ‘basic autonomy’, I
am referring to the simplest systems we know that have all the
required properties for autonomy. The paradigm case is the auto-
poietic cell. Other autonomous systems (e.g. multicellular organisms
or insect colonies) are non-minimal because they are more complex
and not found historically before the appearance of the minimal
versions.
Second, the passage Wheeler quotes from page 160, where I write,
‘minimal autonomy depends on macromolecules but requires that
these macromolecules be organized in a particular way, namely, the
autopoietic way’, occurs in the context of a disagreement with
Dennett (see my response to Dennett above). My point is that, con-
trary to Dennett, replicating macromolecules such as DNA/RNA do
not meet the conditions required for minimal autonomy (in the sense
just indicated), and hence do not meet the conditions required for
agency; they lack the proper organization, namely, the autopoietic
organization.
Finally, it is not the case that ‘the only thing that autopoiesis adds to
the concept of autonomy is the dual materiality of the systemic bound-
ary and systemic mode of existence’. Following Varela (1979), I
define an autonomous system as one in which the constituent pro-
cesses (i) recursively depend on each other for their generation and
their realization as a network; (ii) constitute the system as a unity in
40 E. THOMPSON
whatever domain they exist; and (iii) determine a domain of possible

interactions with the environment (see p. 44). This specification
leaves entirely open the processes that can be interrelated in this way.
What autopoiesis adds to this specification is that the processes are ones
that modulate molecular transformations in the chemical domain. To
repeat a point I made earlier, the crucial feature of such a biochemical
instantiation of autonomy is not the dual materiality per se but the fact
that this materiality realizes a certain organizational boundary. Put
another way, the identity of an autopoietic system cannot be defined
by preservation of the membrane (despite such preservation being
critical for the maintenance of autopoiesis); it has to be defined by
preservation of the network’s organizational boundary, for it is this
boundary that constitutes the crucial interface demarcating the system
from its world.
I come now to the question of whether autonomy and sense-making
require autopoiesis, in the sense of depending constitutively on
autopoiesis. Here I admit to being unsure, for the reasons already
sketched in my reply to Newen. Let me now spell out those reasons in
more detail.
On the one hand, as work on autonomous systems in AI and robot-
ics suggests (see Barandiaran et al., 2009; Froese and Ziemke, 2009),
it seems conceivable that there could be an adaptive self-constituting
system that was not based on autopoietic constituents. For example,
perhaps it is possible to bypass autopoiesis and construct directly a
sensorimotor agent that achieves its autonomy at the level of an adap-
tive and organizationally closed sensorimotor loop.
Notice, however, that for the system to be genuinely autonomous, it
would need to (i) be an individual, in the sense of continually enacting
or bringing forth its own existence in challenging thermodynamic
conditions (where ‘its own existence’ or its individuality is defined by
its topology and organization as a network, not its material or spatial
boundedness per se); (ii) be the active source of its interactions, in the
sense of modulating the parameters of its coupling with the environ-
ment on the basis of its internal (self-organized) activity (‘interactional
asymmetry’); and (iii) generate the norms for those interactions on the
basis of its activity (‘normativity’) (Barandiaran et al., 2009). No
existing robot (e.g. Rodney Brooks ‘Creatures’) meets these criteria.
On the other hand, given these criteria it is not unreasonable to doubt
that they can be achieved without autopoietic constituents. These crite-
ria require, at the sensorimotor level, not just a self-constituting system
embedded in sensorimotor interactions, but a self-constituting system
that can adaptively regulate its sensorimotor interactions. As Froese
and Ziemke (2009) argue in a careful and compelling analysis, such a

system ‘must bring forth its sensors, effectors, and their internal orga-
nizational link (some adaptive mechanism) on the basis of its self-con-
stituting operations. So far, no one has been able to artificially
generate such a system’ (ibid., p. 495). Moreover — and this is the
crucial point for my purposes here — it is hard to see how this require-
ment could be met without something like a metabolism. Put another
way, it is hard to see how this requirement could be met without some-
thing like an autopoietic organization for the constituents that make
up the sensors, effectors, and the adaptive mechanism that links them.
Notice that such an autopoietic organization need not be materially
realized in the same organic way as our terrestrial cells; in that sense,
both the autonomous and autopoietic organizations are composition-
ally plastic. Nevertheless, on the present line of thought, autonomy
would not admit of multiple realizability, in the sense of being
implementable in non-autopoietic mechanisms having causal proper-
ties different from those of autopoiesis.
We thus come back to the second sense of the deep continuity the-
sis, namely that autonomy and sense-making require adaptive
autopoiesis, understood to mean that autonomy and sense-making
depend constitutively on adaptive autopoiesis. If this thesis is true,
then life is necessary is for mind.
This second sense of the deep continuity thesis converges with a
third sense of the thesis, which I will call the existential-phenomen-
ological sense of the deep continuity of life and mind. According to
this thesis, which derives from the writings of Hans Jonas, certain
existential structures of human life or phenomenological structures of
human experience — notably, self/world, freedom/necessity, being/
not-being — are applicable to life itself (pp. 129, 157).
Here is how the convergence between the second and third versions
of deep continuity goes. Jonas traces selfhood and the having of val-
ues, purposes, and norms down to the basic phenomenon of metabo-
lism: ‘to an entity that carries on its existence by way of constant
regenerative activity we impute concern. The minimum concern is to
be, i.e., to carry on being’ (Jonas, 1968). Jonas argues that without
metabolic self-construction, there would be no such thing as the con-
stitution of a meaningful perspective by a system for that system. This
idea provides the existentialist side of the deep continuity thesis:
without constant self-construction or self-creation in a finite, contin-
gent, and challenging environment — in what Ezequiel Di Paolo
(2009) calls ‘precarious conditions’ — there would be no such thing
as subjectivity marked by the polarities of self/world, immanence/
42 E. THOMPSON
transcendence, freedom/necessity, and being/not-being. In this way,

Jonas’s existential-phenomenological analysis of life converges with
the line of thought that suggests that adaptive autonomy and sense-
making depend constitutively on autopoiesis. In short, both lead to the
conclusion that life is necessary for mind.
If mind is necessary for life and life is necessary for mind, then
Wheeler is right that ‘one might as well say that life is identical with
cognition’ (or that life is identical with mind). Like Protevi, Wheeler
thinks this equation results in something like panpsychism, though,
unlike Protevi, who favours panpsychism, Wheeler thinks it ‘doesn’t
so much solve the problem of the genesis of mind as throw a cloak
over the thought that there is a genuine problem to be solved’.
Let me repeat here my response to Protevi. On the one hand, I am
indeed willing to bite the bullet and say that life is mind and mind is
life (though with the reservation that I remain unsure about whether
autonomy requires autopoiesis). On the other hand, if I were writing
Mind in Life today, I would choose a different formulation. I would
say that living is sense-making and that cognition is a kind of
sense-making. But I would do so for the sake of terminological and
conceptual clarity, not in order to mark some clean break in nature. My
aim would be to mark the difference between sense-making as such
(comportment in relation to significance and norms), and the kind of
sense-making that requires intentionality in the proper phenomen-
ological sense — intuitive intentionality (empty and filled intentions in
perception, memory, and imagination), signitive intentionality (pic-
tures, signs, indications), and categorial intentionality (propositional
and conceptual thought).
I am now in position to say something about Wheeler’s discussion
of what ‘enrichment’ means when I write (following Jonas), ‘certain
existential structures of human life are an enriched version of those
constitutive of all life’ (p. 157), or more simply put, that human being
is an enriched version of living being.
In my view, enrichment does not mean that something gets added
from outside life in order to make it mind. Instead, life evolves in such
a way as to transform sense-making, and as this evolution happens the
more complex forms of sense-making reach back, as it were, and
transform the simpler ones. There are many ways to tell this story, but
a Jonas way to tell it would be to say that the self-identity and world-
relatedness of living being evolves into increasingly complex forms
of self-identity and increasingly mediated forms of relating to the
world. On the one hand, the simpler forms are preserved in the more
complex; on the other hand, the more complex forms alter the simpler
ones and make them dependent on the complex ones. For example,
human culture penetrates virtually every aspect of our metabolism, so
that there is no such thing as ‘naked’ human metabolic being inde-
pendent of our cultural ways of living. Enrichment is thus never a
mere addition but always an overall transformation of life and mind.
For this reason, even if it should turn out that life is not necessary for
mind (that autonomy and sense-making do not require autopoiesis), it
would still be the case, contrary to what Wheeler says at the end of his
commentary, that mind would be in life and not simply life in mind.
Wheeler raises one other large issue in his commentary — the rela-
tion of the deep continuity thesis of life and mind to the extended cog-
nition hypothesis. I have discussed this issue elsewhere (Thompson
and Stapleton, 2009), and do not have the space to present a full dis-
cussion here, so I will make only a few points.
Nothing in my view prevents me from allowing that there can be
immanently purposive systems that incorporate elements whose func-
tion is specified extrinsically (see Thompson and Stapleton, 2009).
Think of a prosthetic limb that is incorporated into a person’s ongoing
life. For a system to be immanently purposive it is not necessary that
every element that participates in the system be materially produced
by that system. Furthermore, immanent purposiveness does not mean
that the parts must produce each other in the autopoietic sense; it
means that they must generate and realize themselves as a whole
according to the definition of autonomy. Finally, it remains an open
question whether immanent purposiveness depends constitutively on
autopoiesis, or whether there can be immanent purposiveness without
autopoietic constituents. For these reasons, I see no inconsistency
between my deep continuity view and the extended cognition
hypothesis.
Nevertheless, there are significant tensions between the two
approaches. One tension arises from the way the extended cognition
hypothesis discusses cognition as spatially located versus the way the
enactive approach treats cognition as relational. Wheeler comments
on this issue in note 7, but seems to miss the point. Of course, relations
can exhibit spatiality in the sense that the terms of a relation (e.g. lap-
top, table) are themselves spatially located. But where is the relation
(on top of) itself located? It is something like a category mistake to
think that spatial relations are themselves spatially located in the way
the terms of the relation are. Similarly, the point both Thompson and
Stapleton (2009) and Di Paolo (2009) are making is that it does not
make sense to think of cognition as spatially located in the way that
the ‘vehicles’ enabling cognitive processes are spatially located.
44 E. THOMPSON
Dan Zahavi
Zahavi focuses on my book in relation to the recent debate about the
possibility of ‘naturalizing phenomenology’. He raises two questions.
The first asks whether and how ‘analyses pertaining to subpersonal
processes and mechanisms can possibly influence and enrich phen-
omenological accounts that attempt to do justice to the first-person
perspective and seek to understand the experience in terms of the
meaning it has for the subject’. The second asks ‘how deeply commit-
ted’ I am to transcendental thought. For example, do I ‘endorse some
kind of compatibility between empirical realism and transcendental
idealism’?
The first question Zahavi winds up answering himself. We might
start with a certain phenomenological description and then revise or
enrich this description on the basis of empirical investigation. For
example, a number of traditions in western philosophy and psychology
distinguish between reason and passion, or cognition and emotion.
Some processes are thought to be purely cognitive and others purely
affective. Recent experimental psychology and cognitive neurosci-
ence, however, strongly speak against this view. Many behavioural
and neuroscientific findings indicate that there is no separation
between cognition and emotion: every cognitive process is also affec-
tive, and every brain area traditionally described as cognitive also
belongs to emotion and vice-versa (see Colombetti and Thompson,
2007; Pessoa, 2008; and Chapter Twelve of Mind in Life). These dis-
coveries at neural and behavioural levels can and should provoke a
phenomenological re-examination of experience at the personal level.
What might at first have seemed separate and only instrumentally related
processes in experience (for example, an emotion and a reflective judg-
ment) can be shown to be constitutively interdependent instead.
Nevertheless, I agree with Zahavi that ‘the discovery of a signifi-
cant complexity at a subpersonal level… cannot by itself force us to
refine or revise our phenomenological description. It can only serve as
a motivation for further enquiry’. There is one qualification, however,
that I would add to this remark. Although fMRI or EEG evidence on
its own counts as evidence about only subpersonal neural correlates,
when such evidence is linked to behaviour and self-report it is no lon-
ger purely subpersonal. Instead, it encompasses the personal level as
probed from third-person and second-person perspectives. Motiva-
tions from psychology and neuroscience to refine and revise our
phenomenological descriptions do not come from the strictly subpersonal
level; they come from the way this level is systematically related to the per-
sonal level in experimental investigation.
I also agree with Zahavi that these points are underdeveloped in
Mind in Life. It remains for future work to develop them more system-
atically and with a greater number of examples.
The question of how I see transcendental philosophy in relation to
empirical science is a large issue that I cannot deal with satisfactorily
here, so I will limit myself to a few points.
The foundation of transcendental thinking is rigorous and careful
attention to how things are given to experience. Every claim about
what something is presupposes that the object of this claim is given in
some way to our experience, where ‘experience’ is taken widely to
include thought, and where what counts as a ‘way of experiencing’
cannot be absolutely delimited in advance. Such attention to how
things show up, to the ways they present themselves, necessitates a criti-
cal and reflexive stance toward any cognitive claim, a stance that requires
us to take account of the standpoint of the cognizer in making that claim.
In this way, we are lead back to consider subjectivity and inter-
subjectivity as conditions of knowledge and knowledge production.
This commitment has always been a central component of the
enactive approach (see The Embodied Mind, Chapter One); it is also
what is meant by Maturana and Varela’s statement, ‘everything said is
said by an observer’ (Maturana and Varela, 1987).
Now, this basic commitment to critical reflexivity can be developed
in a variety of different ways and to many different ends. In the case of
transcendental phenomenology, it leads to the analysis of phenomena
such as intentionality, the lived body, intersubjectivity, and time con-
sciousness. In my view, all these analyses — including the ones that
lead Husserl to embrace what he means by transcendental idealism —
are compatible with empirical realism. As Zahavi notes, however, and
as I also maintain, the systematic development of these analyses do
not leave empirical realism unchanged, for they force us to rethink the
concept of nature in ways that can also lead to a transformation of nat-
ural science.
References
Bachman, P.A., et al. (1992) Autocatalytic self-replicating micelles as models for
prebiotic structures, Nature, 357, pp. 57–59.
Barandiaran, X., et al. (2009) Defining agency: Individuality, normativity, asym-
metry and spatio-temporality in action, Adaptive Behavior, 17, pp. 367–386.
Bourgine, P. & Stewart, J. (2004) Autopoiesis and cognition, Artificial Life, 20, pp.
327–345.
46 E. THOMPSON
Christoff, K., et al. (2009) Experience sampling during fMRI reveals default net-
work and executive system contributions to mind wandering, Proceedings of
the National Academy of Sciences USA, 106, pp. 8719–8724.
Clark, A. (2001) Mindware: An Introduction to the Philosophy of Cognitive Sci-
ence, New York and Oxford: Oxford University Press.
Colombetti, G. & Thompson, E. (2007) The feeling body: Toward an enactive
approach to emotion, in Overton, W., Mueller, U. & Newman, J. (eds.) Body in
Mind, Mind in Body: Developmental Perspectives on Embodiment and Con-
sciousness, Hillsdale, NJ: Lawrence Erlbaum Associates.
Cosmelli, D., et al. (2007) Neurodynamical approaches to consciousness, in
Zelazo, P.D., Moscovitch, M. & Thompson, E. (eds.) The Cambridge Handbook
of Consciousness, pp. 731–774, Cambridge: Cambridge University Press.
Dawkins, R. (1995) River out of Eden: A Darwinian View of Life, New York: Basic
Books.
Dennett, D.C. (1991) Consciousness Explained, Boston, MA: Little Brown.
Dennett, D.C. (1992) Review of F.J. Varela, E. Thompson and E. Rosch, The
Embodied Mind: Cognitive Science and Human Experience, and G. Edelman,
Bright Air, Brilliant Fire: On the Matter of Mind, New Scientist, 13 June, pp.
48–49.
Dennett, D.C. (1993) Review of F.J. Varela, E. Thompson and E. Rosch, The
Embodied Mind: Cognitive Science and Human Experience, American Journal
of Psychology, 106, pp. 121–126.
Di Paolo, E.A. (2009) Extended life, Topoi, 28, pp. 9–21.
Farb, N.A.S., et al. (2007) Attending to the present: Mindfulness meditation
reveals distinct neural modes of self-reference, Social Cognitive Affective Neu-
roscience, 2, pp. 313–322.
Froese, T. & Ziemke, T. (2009) Enactive artificial intelligence: Investigating the
systemic organization of life and mind, Artificial Intelligence, 173, pp.
466–500.
Gallagher, S. & Meltzoff, A. (1996) The earliest sense of self and others:
Merleau-Ponty and recent developmental studies, Philosophical Psychology, 9,
pp. 213–236.
Jonas, H. (1966) The Phenomenon of Life: Toward a Philosophical Biology, Chi-
cago, IL: University of Chicago Press.
Jonas, H. (1968) Biological foundations of individuality, International Philosoph-
ical Quarterly, 8, pp. 231–251.
Kelso, J.A.S. (1995) Dynamic Patterns: The Self-Organization of Brain and
Behavior, Cambridge, MA: MIT Press/A Bradford Book.
Kelso, J.A.S. & Kay, B.A. (1987) Information and control: A macroscopic analysis
of perception-action coupling, in Heuer, H. & Sanders, A.F. (eds.) Perspectives
on Perception and Action, pp. 3–32, Hillsdale, NJ: Lawrence Erlbaum
Associates.
Luisi, P.L. (2010) The Emergence of Life: From Chemical Origins to Synthetic
Biology, Cambridge: Cambridge University Press.
Lutz, A., et al. (2002) Guiding the study of brain dynamics by using first-person
data: Synchrony patterns correlate with ongoing conscious states during a sim-
ple visual task, Proceedings of the National Academy of Sciences USA, 99, pp.
1586–1591.
Lutz, A., et al. (2008a) Attention regulation and monitoring in meditation, Trends
in Cognitive Sciences, 12, pp. 163–169.
Lutz, A., et al. (2008b) Regulation of the neural circuitry of emotion by compas-
sion meditation: Effects of meditative expertise, PLoS ONE, 3 (3), e1897,
DOI:10.1371/journal.pone
Maturana, H.R. & Varela, F.J. (1987) The Tree of Knowledge: The Biological
Roots of Human Understanding, Boston, MA: Shambala Press/New Science
Library.
McMullin, B. (2004) Thirty years of computational autopoiesis: A review, Artifi-
cial Life, 10, pp. 277–295.
Merleau-Ponty, M. (1963) The Structure of Behavior, Fisher, A. (trans.), Pitts-
burgh, PA: Dusquene University Press.
Moreno, A. & Barandiaran, X. (2004) A naturalized account of the inside-outside
dichotomy, Philosophica, 73, pp. 11–26.
Nielsen, L. & Kaszniak, A. (2006) Awareness of subtle emotional feelings: A com-
parison of long-term meditators and nonmeditators, Emotion, 6, pp. 392–405.
Noë, A. (2004) Action in Perception, Cambridge, MA: MIT Press.
Noë, A. (2005) Real presence, Philosophical Topics, 33, pp. 235–264.
Pessoa, L. (2008) On the relationship between emotion and cognition, Nature
Reviews Neuroscience, 9, pp. 148–158.
Petitmengin, C., et al. (2007) Anticipating seizure: Pre-reflective experience at the
center of neurophenomenology, Consciousness and Cognition, 16, pp.
746–764.
Piccinini, G. (2010) How to improve on heterophenomenology: The self-measure-
ment of first-person data, Journal of Consciousness Studies, 17 (3–4), pp.
84–106.
Protevi, J. (2010) Larval Subjects, Autonomous Systems, and E. Coli Chemotaxis,
[Online], http://www.protevi.com/john/Deleuze_enaction.pdf
Ruiz-Mirazo, K. & Moreno, A. (2004) Basic autonomy as a fundamental step in
the synthesis of life, Artificial Life, 10, pp. 235–259.
Sartre, J.-P. (2004) The Imaginary: A Phenomenological Psychology of the Imagi-
nation, Webber, J. (trans.), London and New York: Routledge Press.
Shapiro, L. (2004) The Mind Incarnate, Cambridge, MA: MIT Press/A Bradford
Book.
Sober, E. (1984) The Nature of Selection, Cambridge, MA: MIT Press/A Bradford
Book.
Sokolowski, R. (2000) An Introduction to Phenomenology, Cambridge: Cam-
bridge University Press.
Thompson, E. (1992) Novel colours, Philosophical Studies, 68, pp. 321–349.
Thompson, E. (1995) Colour Vision: A Study in Cognitive Science and the Philoso-
phy of Perception, London: Routledge Press.
Thompson, E., et al. (1992) Ways of coloring: Comparative color vision as a case
study for cognitive science, Behavioral and Brain Sciences, 15, pp. 1–26.
Thompson, E. & Stapleton, M. (2009) Making sense of sense-making: Reflections
on enactive and extended mind theories, Topoi, 28, pp. 23–30.
Varela, F.J., Thompson, E. & Rosch, E. (1991) The Embodied Mind: Cognitive
Science and Human Experience, Cambridge, MA: MIT Press.
Weber, B.H. & Depew, D.J. (1995) Darwinism Evolving: Systems Dynamics and
the Genealogy of Natural Selection, Cambridge, MA: MIT Press.
Weber, B.H. & Depew, D.J. (1996) Natural selection and self-organization, Biol-
ogy and Philosophy, 11, pp. 33–65.
Welton, D. (2011) Can a top-down phenomenology of intentional consciousness
be integrated with a bottom-up phenomenology of biological systems?, Philos-
ophy Today, in press.
Zahavi, D. (2005) Subjectivity and Selfhood: Investigating the First-Person Per-
spective, Cambridge, MA: MIT Press/A Bradford Book.

E. Thompson

Uploaded by

Copyright:

Available Formats

E. Thompson

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

E. Thompson

Uploaded by

Copyright:

Available Formats

Evan Thompson

Journal of Consciousness Studies, 18, No. 5–6, 2011, pp. ??–??

takes to be criticisms of my view whereas they are actually points I

an air of noblesse oblige, circumscribe its force by claiming to have

functions for converting input instructions into output products. Of

present in Mind in Life goes well with the statistical interpretation of

specific model that makes predictions to be tested against other mod-

conflates that perspective with the phenomena themselves. Dennett’s

2. Developmental Systems Theory

genes is misguided: genes are not units of information in any sense of

version is exactly the dualism Dennett expresses in the passage I quote

Genes are pure information — information that can be encoded,

I agree with Dennett that AI systems can in principle be designed to

projects a faulty conception of agency onto entities that do not meet

4. Phenomenology and Heterophenomenology

sense that it is about or directed toward an object, and every intentional

basis, so that more precise correlations can be established between the

For some examples of first-person/second-person methods being

Lucia Foglia and Rick Grush

sensorimotor processes in mental imagery (see pp. 295–6, 298–9).

natural selection is conceived as an outside force acting on agents con-

1. Autopoiesis, Adaptivity, and Cognition

(2005) calls ‘adaptivity’, namely, being able to modulate the auto-

2. Does Autopoiesis Plus Adaptivity Entail Sense-Making?

states. Without such internal representations, he maintains, we have no

3. The Deep Continuity of Mind and Life

According to the strong reading, ‘the principles of mind are essen-

concerned perspective on the world that comes from having con-

5. Can Life Be Known Only by Life?

First, the transcendental framework I am working with in Mind in

he read Siewert’s commentary, which ably elaborates my view that

7. Enaction and Consciousness

was already implicit in his earlier work on autopoiesis with

to respond, so I will focus on what he calls the ‘question of panpsych-

its momentary expression. Possessing internal unity inscribed in a seg-

passages back and forth between those orders that we conceptualize —

functions occurring in our world could not occur in a world in which

Siewert’s own presentation of the embodied-consciousness view

Robert Van Gulick

biosemantics, and teleofunctionalism advance a kind of naturalism

(6) Thus, life is autopoiesis plus cognition (any living system is an

boundary and the systemic mode of existence’ is necessary for auton-

in a relevantly different way (Shapiro, 2004). We can thus distinguish

whatever domain they exist; and (iii) determine a domain of possible

and Ziemke (2009) argue in a careful and compelling analysis, such a

transcendence, freedom/necessity, and being/not-being. In this way,

You might also like