Thursday, May 24, 2018

An Argument Against Every General Theory of Consciousness

As a philosophical expert on theories of consciousness, I try to keep abreast of the most promising recent theories. I also sometimes receive unsolicited emails from scholars who have developed a theory that they believe deserves attention. It's fun to see the latest cleverness, and it's my job to do so, but I always know in advance that I won't be convinced.

I'd like to hope that it's not just that I'm a dogmatic skeptic about general theories of consciousness. In "The Crazyist Metaphysics of Mind", I argue that our epistemic tools for evaluating general theories of consciousness are, for the foreseeable future, too flimsy for the task, since all evaluations of such theories must be grounded in some combination of dubious (typically question-begging) scientific theory, dubious commonsense judgment (shaped by our limited social and evolutionary history), and broad criteria of general theoretical virtue like simplicity or elegance (typically indecisive among theories that are live competitors).

Today, let me try another angle. Ultimately, it's a version of my question-beggingness complaint, but more specific.

Premise 1: There is no currently available decisive argument against panpsychism, the view that everything is conscious, even very simple things, like solitary hydrogen ions in deep space. Panpsychism is, of course, bizarrely contrary to common sense, but (as I also argue in The Crazyist Metaphysics of Mind) all well-developed general theories of consciousness will have some features that are bizarrely contrary to common sense, so although violation of common sense is a cost that creates an explanatory burden, it is not an insurmountable theory-defeater. Among prominent researchers who defend panpsychism or at least treat seriously a view in the neighborhood of panpsychism are Giulio Tononi, David Chalmers, Galen Strawson, and Philip Goff.

There are at least three reasons to take panpsychism seriously. (1.) If, as some have argued, consciousness is a fundamental feature of the world, or a property not reducible to other properties, it would be unsurprising if such a feature were approximately as widespread as other fundamental features such as mass and charge. (2.) Considering the complexity of our experience (e.g., our visual experience) and the plausibly similar complexity of the experience of other organisms with sophisticated sensory systems, one might find oneself on a slippery slope toward thinking that the least complex experience would be possessed by very simple entities indeed (see Chalmers 1996, p 293-7, for a nice exposition of this argument). (3.) Despite my qualms about Integrated Information Theory, there's an attractive theoretical elegance to the idea that consciousness arises from the integration of information, and thus that very simple systems that integrate just a tiny bit of information will correspondingly have just a tiny bit of consciousness.

Premise 2: There is no currently available decisive argument against theories of consciousness that require sophisticated self-representation of the sort that is likely to be absent from entities that lack theories of mind. On extreme versions of this view, even dogs and infants might not have conscious experience. (Again, highly contrary to common sense, but!) Among prominent researchers who have taken such a view seriously are Daniel Dennett and Peter Carruthers (though recently Carruthers has suggested that there might be no fact of the matter about the phenomenal consciousness, or not, of non-human animals).

There are at least three reasons to take seriously such a restrictive view of consciousness: (1.) If one wants to exit the slippery slope to panpsychism, one possibly attractive place to do so is at the gap between creatures who are capable of explicitly representing their own mental states and those that cannot do so. (2.) Consciousness, as was noted by Franz Brentano (and recently emphasized by David Rosenthal, Uriah Kriegel, and others), might plausibly always involve some sort of self-awareness of the fact that one is conscious -- apparently a moderately sophisticated self-representational capacity of some sort. (3.) There's a theoretical elegance to self-representational theories of consciousness. If consciousness doesn't just always arise when information is integrated in a system, an attractive explanation of what else is needed is some sort of sophisticated ability of a system to represent its own representational states.

Now you might understandably think that either panpsychism or a human-only views of consciousness is so extreme that we can be epistemically justified in confidently rejecting one or the other. If so, we can run the argument with weaker versions of Premise 1 and/or Premise 2:

Premise 1a (weaker): There is no currently available decisive argument against theories of consciousness that treat consciousness as very widespread, including perhaps in organisms with fairly small and simple brains, or in some near-future AI systems.

Premise 2a (weaker): There is no currently available decisive argument against theories of consciousness that treat consciousness as narrowly restricted to a class of fairly sophisticated entities, perhaps only mammals and birds and similar organisms capable of complex, flexible learning, and no AI systems in the foreseeable future.

Premise 3: All general theories of consciousness commit to the falsity of either Premise 1, Premise 2, or both (alternatively Premise 1a, Premise 2a, or both). If they do not so commit, then they aren't general theories of consciousness, though they may of course be perfectly fine narrow theories of consciousness, e.g., theories of consciousness as it happens to arise in human beings. (I've got a parallel argument against general theories of consciousness even as they apply just to human beings, based on considerations from Schwitzgebel 2011, ch. 6, but not today.)

Therefore, all general theories of consciousness commit to the falsity of some view against which there is no currently available decisive argument. They thereby commit beyond the evidence. They must either assume, or accept on only indecisive evidence, either the falsity of panpsychism, or the falsity of sophisticated self-representational views of consciousness, or both. In other words, they inevitably beg the question against, or at best indecisively argue against, some views we cannot yet justifiably reject.

Still, go ahead and build your theory of consciousness. You might even succeed in building the true theory of consciousness, if it isn't yet out there! Science and philosophy needs bold theoretical adventurers. But if a skeptic on the sidelines remains unconvinced, thinking that you have not convincingly dispatched some possible alternative approaches, the skeptic will probably be right.

ETA: In order to constitute an argument against a candidate theory, as opposed to merely an objection to such theories, perhaps I need to put some weight on the positive arguments in favor of views of consciousness that conflict with the theory being defended. Thanks to David Chalmers and Francois Kammerer on Facebook for pushing me on this point.

[image source]

Tuesday, May 15, 2018

"What Is It Like" Interview Now Freely Available

... here.

It's a fairly long read (about 8000 words), but I gave a lot of thought to Cliff's questions, and I hope the result is both interesting and revealing.

Read it, and learn about:

* my unconventional parents, including how we celebrated Christmas and my part-time work as a chambermaid;

* the peculiar story of how I once found my lost wallet;

* sneaking into György Gergely's cognitive development class at Berkeley, and how I think "Stanford school" philosophy of science should inform philosophy of mind;

* what four-year-olds and philosophers have in common;

* why blogging is the ideal form for philosophy;

* and much more!

Also please consider supporting Cliff Sosis's "What Is It Like to Be a Philosopher?" interviews by funding him on PayPal or Patreon.

Friday, May 11, 2018

Is C-3PO Alive?

by Will Swanson and Eric Schwitzgebel

Droids—especially R2-D2, C-3PO, and BB-8—propel the plot of the Star Wars movies. A chance encounter between R2-D2 and Luke Skywalker in “Episode IV: A New Hope” starts Luke on his fateful path to joining the rebel forces, becoming a Jedi, and meeting his father. More recently, BB-8 plays a similar role for Rey in “Episode VII: The Force Awakens”. But the droids are more than convenient plot devices. They are full-blooded characters, with their own quirks, goals, preferences, and vulnerabilities. The droids face the same existential threats as anyone else; and most of us still squirm in our theater seats on their behalf when danger looms.

Our response to the Star Wars droids relies on the tacit assumption that they are living lives -- lives that can be improved or worsened, sustained or lost. This raises the question: Is C-3PO alive? Or more precisely, if we someday built a robot like C-3PO, would it be alive?

In a way, it seems obvious that, yes, C-3PO is alive. Vaporizing him would be murder! One could have a funeral over his remains, reminiscing about all the good things he did in his lifetime.

But of course, the experts on life are biologists. And if you look at standard biology textbook descriptions of the characteristics of life (e.g., here), it looks like robots wouldn’t qualify. They don’t grow or reproduce, or share common descent with other living organisms. They don’t contain organic molecules like nucleic acid. They don’t have biological cells. They don’t seem to have arisen from a Darwinian evolutionary process. Few people (and probably fewer professional biologists) would say that a Roomba vacuum cleaner is alive, except in some kind of metaphorical sense; and in these respects, C-3PO is similar, despite being more complicated – just as we are similar to but more complicated than microscopic worms. The science covering C-3PO is not biology, but robotics.

Despite what looks like bad news for C-3PO from biology textbook definitions of life, on closer consideration we should reject the biology-textbook-list approach to robot cases. Our attitude toward these lists should probably be closer to the capacious attitude typical of astrobiologists (e.g., Benner 2010). If we’re considering what “life” is, really, in the broad, philosophical way that we do when considering the possibility of alien life, the standard lists start to look very Earth-bound and chauvinistic.

  • Common descent. Unless we wish to exclude the possibility of life originating independently on other planets, we should not treat common descent as a necessary condition for life.
  • Organic molecules. If we allow for life to arise independently on other planets, we should also be wary of expecting the resultant life to closely resemble biological life on Earth. We should not require the presence of organic molecules like nucleic acids.
  • Reproduction. While it is true that biological living things tend to reproduce when given the opportunity, reproduction is far from necessary for life. Consider the mule, the sterile worker ant, and the deliberately childless human. Nor should we require that life forms originate from reproductive processes. If life began without reproduction once, it can begin again, perhaps many times over!
  • Participation in Darwinian processes. Explanations invoking evolution by natural selection have revealed many of nature’s secrets. Nevertheless, evolution is not a locally defined property of living individuals. It refers to the processes that shape individuals over generations. It’s unclear why belonging to a group that has undergone Darwinian selection in the past should matter to whether an individual, considered now, is alive.
  • Growth. Depending on the sense of growth in question, robots may or may not grow. If growth means nothing more than change over time in accordance an internal protocol, then at least some robots, learning ones, are able to grow. If growth means simply getting radically bigger (or developing from a small seed or embryo into a large adult), then requiring growth risks excluding or marginalizing many organisms that are uncontroversially alive, such as bacteria that reproduce by fission.
  • Other list members -- self-maintenance (if a robot can charge its own battery...), having heterogeneous organized parts, and responding to stimuli -- pose no challenge to the idea that robots are alive.

    What about metabolism? Perhaps this is an essential feature of life. Do robots do this?

    Tweaking a suggestion from Peter Godfrey-Smith (2013), a first pass on a definition for metabolism is the cooperation of diverse parts within an organism (implicitly, a thing that meets other criteria for life) to use energy and other resources to maintain the structure of the organism. If “maintaining structure” amounts to maintaining operational readiness, then this definition provides no reason to deny metabolism to robots, especially robots that do things like auto-update, repel virus programs, and draw from external energy resources as needed. If “maintaining structure” refers specifically to the upkeep required to keep a physical body from degrading, then most simple robots would be excluded, but C-3PO would still qualify, if he can polish his head and order a replacement arm.

    Even so, this second approach might define metabolism too narrowly. Defining metabolism in terms of maintenance in a narrow sense, after all, cannot accommodate the other ends to which organisms put energy in coordinated, non-accidental ways. Consider growth and development. The caterpillar’s metamorphosis hardly counts as maintenance of structure in any straightforward sense, yet we should count the energy transformations needed to effect that change as part of the caterpillar’s metabolism. More strikingly, living things often use energy to undermine their structure: think of cells undergoing programmed cell death or humans committing suicide.

    We can accommodate these cases by broadening the definition of metabolism to encompass any coordinated use of energy within a living thing to achieve the ends of that living thing. On this broader definition, there is no reason to deny metabolism to robots.

    With all of this in mind, we think it’s not unreasonable to stick with our gut intuition that C-3PO is alive. What is essential to being a living thing is not so much one’s biological history or composition by organic molecules, but rather the use of internal or environmental resources to accomplish the functional aims of the system.

    How sophisticated does a system need to be to qualify as living, by these standards? Should we maybe say that even a Roomba is alive, after all? In a series of entertaining experiments, Kate Darling has shown that ordinary people are often quite reluctant to smash up cute robots. Despite Darling’s own expressed view that such robots aren’t alive, maybe part of what is holding people back is some of the same thing that holds some of us back from wanting to crush spiders -- a kind of emotional reverence for forms of life.

    A darling robot:

    Philosophers have grown used to functionalism about mind -- that is, they seem generally willing to accept or at least take seriously the possibility that consciousness might be realized in non-biological substrates. Nevertheless, functionalism about life is less readily accepted. Perhaps philosophical reflection about the possibility of robotic life can help us recognize that our concern over the lives and deaths of our favorite robot characters may be perfectly justified.

    [C-3PO image source] [Pleo image source]

    Wednesday, May 09, 2018

    What Is It Like Interview

    Cliff Sosis has interviewed me for his wonderful What Is It Like to Be a Philosopher series. For the first week, it is only available to Silver level patrons on Patreon. Next week, he'll release it more broadly.

    Please consider supporting Cliff's work. Cliff's interviews knit together questions about philosophy with questions about childhood, family, hobbies, passions, formative experiences -- giving the reader a fuller sense of the whole person than one generally sees. Check out his interviews with Josh Knobe, Kwame Appiah, David Chalmers, Sally Haslanger, Peter Singer, and Kate Manne, for example.


    In this interview Eric Schwitzgebel, professor of philosophy at University of California Riverside, and I discuss his father’s collaboration with Timothy Leary, his uncle who invented ankle monitors, pretending to be a mannequin for his father’s class, Christmas with an electric blue Buddha, his mother’s anti-theist views, being a chambermaid and skiing, writing plays, Rosencrantz and Guildenstern are Dead, T.S. Eliot, Stephen Jay Gould, Stanford and Kuhn, Feyerabend and Zhuangzi, disliking analytic philosophy, moving from Deleuze and Derrida to Hume and Dretske, living in a hippy co-op, wearing a dress as a political statement, memorizing Sylvia Plath, Dretske and Dupré, the Gourmet Report, working with Elisabeth Lloyd and John Searle, the allegations against Searle, the grad culture at Berkeley, love and death, the Bay Area Philosophy of Science reading group, Peter Godfrey-Smith, Stanford School philosophy of science, Bayes or Bust?, sneaking into György Gergely’s class, Alison Gopnik’s generosity, meeting his wife via a classified ad, the job market in 97, landing a job at U.C. Riverside where he is to this day, how the department has changed and he has changed as a teacher, his class on Evil, his work on the moral behavior of ethics professors, The Splintered Mind and philosophy on the internet, his theory of jerks, Brian Weatherson, experimental philosophy, our philosophical blind spots, his writing routine and process, work-life balance, My Dinner with Andre, election night 2008 versus election night 2016, and his last meal…

    Friday, May 04, 2018

    The Rise and Fall of Philosophical Jargon

    In 2010, I defined a discussion arc, in philosophy, as a curve displaying the relative frequency at which a term or phrase appears among the abstracts of philosophical articles. Since then, I've done a few analyses using discussion arcs, mostly searching for philosophers' names (here, here, here, here, here).

    Today I thought it would be fun to look at philosophical jargon -- its rise and fall, how much it varies over time, and as a measure of what is hot. Maybe you'll find it fun too!

    I rely on abstract searches in The Philosophers Index. NGram is nifty, but it doesn't do well with trends specifically in academic philosophy (see here). JStor is interesting too, but it's a limited range of journals, especially for articles less than six years old.

    First, I constructed a representative universe of articles (limiting the search to journal articles only): In this case, I searched for "the" in the abstracts, in five year intervals from 1940-present, except merging 1940-1949 for smoothness and a large enough sample. Then I searched for target terms in abstracts in the same five-year intervals. I constructed the curves by dividing the number of occurrences of the target term by the number of articles in the representative universe in each period.

    Some topics and terms are perennial, rising and falling a little, but not in any dramatic way. Others increase or decrease fairly steadily without a clear peak (allowing for some noisiness especially in the early data). For example, here are the are the arcs for happiness, Kant*, and skepticism or scepticism (all three fairly steady), evolution* and democra* (both rising), and induction (falling):

    (The asterisk is a truncation symbol; the y-axis is percentage of all abstracts containing the word.)

    [apologies for blurry image; click to enlarge and clarify]

    (You thought happiness was more important to philosophers than Kant? Wrong!)

    One way to measure how trendy or steady a topic is, is the ratio of percentage of discussion in its peak period, compared to its average discussion. Exactly equal discussion over the whole period would yield a ratio of 1:1. Fairly steady discussion with some noise would be 1 to 1.5. Topics that rise and fall more substantially would be proportionally higher. Call this the Max/Average Ratio. For the six topics above, the Max/Average Ratios are:

  • Kant*: 1.3
  • happiness: 1.4
  • skepticism or scepticism: 1.4
  • evolution*: 1.5
  • democra*: 2.0
  • induction: 3.0
  • Evolution*, though it approximately triples over the period, has a Max/Average Ratio not too far from one. Democra* rises from a substantially lower level of discussion than does evolution* and has a higher Max/Average Ratio. Induction crashes down to about a sixth of its initial level of discussion (0.174% in the first four periods to 0.028% in the final two) -- hence its moderately large ratio.

    Now let's consider some jargon terms that more clearly reflect hot topics.

    Since the scale is logarithmic, periods of zero incidence are not charted. Also remember that logarithmic scales visually compress peaks relative to linear scales. For example, though the decline of "language of thought" is not so visually striking, usage was in fact about seven times as much at peak as it is now.

    "Grue" was introduced by Nelson Goodman to describe a puzzle about how we know the future in his "New Riddle of Induction" in 1955. As you can see from the chart, discussion peaked 10-15 years later, in the late 1960s. "The original position" was introduced by John Rawls in his 1971 A Theory of Justice, describing part of an idealized decision-making process, and discussion of it peaked in the late 1980s, about 15 years later. "Supervenience" has a murkier origin, but was popularized in philosophy by R.M. Hare in 1952, to talk about how one set of properties might covary with another (for example the moral and the physical). Discussion peaked about 40 years later in the early 1990s. Hilary Putnam introduced "Twin Earth" (a planet with XYZ instead of water) in a thought experiment in 1975, and discussion peaked 15-20 years later in the early 1990s. "Radical interpretation" was introduced by Donald Davidson in the early 1970s, peaking 15 years later in the late 1980s. Finally, the "language of thought" was introduced by Jerry Fodor in his 1975 book of the same title, peaking 15-20 years later in the early 1990s.

    With the exception of supervenience -- maybe partly because the concept took some time to transition from ethics to the metaphysics of mind -- the pattern is remarkably consistent, with peaks about 15-20 years after a famous introduction event. This pacing reminds me of my earlier research suggesting that individual philosophers tend to receive peak discussion around ages 55-70, despite producing their most influential work on average about 20 years earlier (NB: the two data sets are only partly comparable, but I'm pretty sure the generalization is roughly true anyway). This is the pace of philosophy.

    For these terms and phrases, the Max/Average Ratios are a bit higher than for the rising and falling topics sampled above:

  • superven*: 2.7
  • "radical interpretation": 3.4
  • "Twin Earth: 3.9
  • "the original position": 4.2
  • "language of thought": 4.6
  • grue: 5.3
  • The Max/Average Ratio, of course, doesn't really capture rising and falling; and the ratio will be inflated for more recently introduced terms, assuming virtually zero incidence before introduction.

    For a better measure of peakiness, we can examine the ratio of the maximum discussion to the average discussion in the first two and final two time periods. To avoid infinite peakiness and overstating the peakiness of rare terms, I'll assume a floor level of discussion of .01% in any period. Call this Peakiness with a capital P. Five of the six topics in the first group have a Peakiness between 1.3 and 2.0, and evolution* has 2.9.

    In contrast:

  • superven*: 3.8
  • "Twin Earth": 4.7
  • "radical interpretation": 6.3
  • "the original position": 8.4
  • "language of thought": 8.8
  • grue: 50.6
  • Grue was hot! Although its peak discussion was about the same as superven*, it has crashed far worse -- or at least it has, so far. If we had a longer time period to play with, we could try to make the analyses more temporally stable by sampling a window of 50 years before and after peak, thus giving superven* more of a fair chance to finish its crash, as it were.

    Okay, how about newer jargon? Let's try a few. I guess first I should say that jargon is wonderful and valuable, and I actually love the grue and Twin Earth thought experiments. Also some jargon becomes a permanent part of the discipline -- such as "dualism" and "secondary qualities". Maybe grue and Twin Earth will also prove in the long run to be permanent valuable contributions to the philosopher's toolbox, just in a lower-key way than back when they were hot topics. I don't really mean "jargon" as a term of abuse.

    In 1974, Robert Kirk introduced "zombies" in the philosophical sense (entities physically indistinguishable from people but with no conscious experience), but usage didn't really take off until they got discussion in several articles in Journal of Consciousness Studies in 1995 and in David Chalmers' influential The Conscious Mind in 1996. Contrary to popular rumor, the zombie doesn't appear to be dead quite yet. "Epistemic injustice" was introduced by Miranda Fricker in the late 1990s and early 2000s. "Virtue epistemology" was introduced by Ernest Sosa in the early 1990s. Fictionalism has a longer and more complicated history in logic and philosophy of math.

    The "explanatory gap" between physical or functional explanations and subjective conscious experience was introduced by Joseph Levine in 1983, but doesn't hit the abstracts until some papers of his in the early 1990s. "Experimental philosophy" in its earlier uses refers to the early modern scientific work of Newton, Boyle, and others. It's recent usage to refer to experimental work on folk intuitions about philosophical scenarios hits the abstracts all at once with five papers in 2007. Consistently with my twenty-year hypothesis, of these, "explanatory gap" is the only one that shows signs of being past its peak (despite hopes expressed by some of my Facebook friends). Maybe fictionalis* is running longer.

    Okay, I can't resist trying a few more discussion arcs, just to see how they play out.

    "Possible worlds" goes back at least to Leibniz, but its first appearance in the abstracts was by Saul Kripke in a 1959 article. It peaks at 2.0% in the late 1960s, but has enduring popularity (currently 0.4%). "Sense data" as the objects of perception was introduced in the early 20th century by G.E. Moore and Bertrand Russell and has lots of discussion in the beginning of this dataset (1.7%), crashing down to levels appropriate to a historical relic (0.02%). "Qualia" has a couple occurrences early in the abstracts and traces back to C.S. Peirce in the 19th century, then hits the abstracts again with an article by Sydney Shoemaker in 1975, peaking in the late 1990s.

    Supererogation (morally good acts beyond what is morally required) entered modern discussion in the late 1950s and early 1960s (first hitting the abstracts with Joel Feinberg in 1961), then peaked in the late 1980s -- but it looks like it might be staging a comeback? Wittgenstein introduced the idea of the "language game" in his posthumous Philosophical Investigations (1953), with discussion peaking in the late 1970s. Thomas Nagel introduced "moral luck" in a classic 1979 article, and although it peaked in the late 2000s, it hasn't yet declined much from that peak.

    Possible worlds has the highest Peakiness of the lot -- though nothing like grue -- at 8.4. "Language game" is next at 5.1. The rest aren't very Peaky, ranging from 2.1 to 2.9.

    I've spent more hours this week doing this than I probably should have, given all my other commitments -- but there's something almost meditative about data-gathering, and the arcs yield a perspective I appreciate on the historical brevity of philosophical trends.

    Tuesday, May 01, 2018

    Please Rate My Blog Posts for Inclusion in My Next Book

    I'm working away on selecting and revising blog posts and op-eds for my next book. Readers' feedback has been very helpful in narrowing down to just the most memorable and interesting posts! My final poll, 22 selected posts on philosophical method and the sociology of philosophy is live today.

    As with all the other polls, this poll contains links to the original posts so so you can refresh your memory if you want. But there's no need to rate all of the listed posts! Even if you just remember one or two that you like, it would be useful for me to know that.

    Below are all seven polls.

    Polls 3, 5, and 6 have low response rates. It would be terrific if you could click through and rate a few posts that you like or remember, from one or more of those polls.


  • 1. moral psychology
  • 2. technology
  • 3. belief, desire, and self-knowledge
  • 4. culture and humor
  • 5. cosmology, skepticism, and weird minds
  • 6. consciousness
  • 7. philosophical method and the sociology of philosophy
  • (new as of today)