Interview by Richard Marshall.

Andy Clarkis currently professor of philosophy at the University of Edinburgh, UK. His research centers on artificial intelligence, artificial life, embodied cognition, philosophy of mind, technology, and culture. Here he shoots from the hip on the extended mind, flip flopping, Searles' mind-brain thesis, extended consciousness, lazy brains and rich world models, Terminators and Star Trek, something about Chinese rooms, cameras and understanding, Chalmers and dualism, a new paradigm of the mind, consciousness and the self, and playing ethical catch-up with the singularity. Va-va-Voom!

3:AM:What made you become a philosopher?

Andy Clark:Kinda random really – I was thrown out of my French Lit degree for arrogance. I was taking Philosophy as an ‘easy’ minor and I found myself strangely good at logic, and into A. I....So oddly I stayed there.

3:AM:You have been a key figure in the development of the extended mind. Could you sketch for us the salient points about this rather strange thesis that says that when I use a notebook in a certain proscribed systemic way the notebook is not merely a useful helpmate for my thinking but is an extension of my mind. Have I got that right, that in theses cases, such as when the character Otto uses his notebook the notebook is part of his mind just as his brain is?

AC:Yes - The core idea is that what matters is not where stuff is encoded, or in what medium, but the uses to which it can readily be put. Poise is everything. Just as it doesn’t really matter, when working online, whether some piece of information is stored on your hard drive or in the cloud, as long as it’s usually ready for access when the need is there.

I think it also helps to consider a documented real- world case such as that of Patrick Jones. Jones suffered severe memory impairments (much like those of the lead character in the film Memento) as a result of repeated traumatic brain injury. Yet he lives a surprisingly normal life as a working catholic deacon in Colorado Springs. To do so, Jones relies upon a complex combination of software packages and a smartphone. Courtesy of these off- the-shelf packages and devices Jones has been able to create massive webs of interlinked notes and pointers that allow the saving, searching, retrieving, and diagramming of his own contacts, thoughts, meetings, decisions, and interactions. It is only in virtue of this whole up-and-running web of structure that he able to recall who he has spoken with, what was decided, what projects are currently live, and so on. Yet he carries through complex long-term projects of pastoral care with incredible skill, optimism, and good humour. Patrick, the person, is now built of both biological and non-biological parts, some of the latter not even being attached to his biological body. If you were to hack into and destroy his data trails, that feels to me like a crime against the person, not merely a crime against his cyber-property. But If Patrick Jones looks like a possible case of an ‘extended mind’, then so, surely, are we. Our own pens ,papers, notebooks, diaries, smartphones and (increasingly distributed) data-trails have evolved to complement our brute biological profile too.

3:AM:How do you prevent everything becoming the mind if it’s useful to my thinking? For instance, why aren’t the books in my room my mind – after all, I consult them to find things out when stuck? Or the internet etc? If the skin isn’t the barrier aren’t you threatened by the difficulty of restraining the mind?

AC:Not really. It’s all about reliable access and sufficient meta-knowledge. So you do need to know what kind of thing you know.......just like when using your smartphone you know you in effect know those ‘phone numbers. Just like, when using bio-memory, you know you know your mother’s maiden name even before it comes back to you.

3:AM:You argue that we can flip from something (such as Otto’s notebook) being something accessed by perception to being an act of memory. So the notebook is at some time not an extended mind and at other times it is. If I’ve got that right, how does this work and why doesn’t this flip flop count against the theory?

AC:Good question. I think the right answer is that the route through perception is actually a feature not a bug. My co-author on the extended mind paper, Dave Chalmers, defends this in a new paper coming soon in a volume from OUP edited by Irvine, Colombo, and Stapleton (three of my ex PhD students). It’s called Clark and his Critics....a scarily end-of-working-life kind of title...

Anyway, in the forthcoming paper Chalmers’ argues that the interesting thesis at issue in these debates is that some of the machinery of mind can safely be located beyond the intuitive bounds of perception and action.. Dave invites me to sign up for something like this this as the official statement of the core thesis. I hereby do so. Thanks Dave.

3:AM:John Searle says that the mind is the brain. Doesn’t this approach wreck that thought?


3:AM:The Terminator retrieves information by reading off information from a screen. If the flip thesis works then we can concede that the Terminator has a mind (or at least we needn’t discount it just on the grounds that he reads the information). But do you think that out of your extended mind thesis we might have things like extended desire, extended emotion, extended perception?

AC:Yes – for non-conscious emotion etc. But I’m not sure about extended consciousness. Probably I have to embrace this, as Giovanna Colombetti argues. But I
am still on the fence.

3:AM:Might Terminator or Datain Star Trek have feelings after all?

AC:Yes. Feelings are one thing. Knowing you have feelings is another. Many episodes of Star Trek explore this potential tension, boldly going where many philosophers and cognitive scientists have gone before.

3:AM:I suppose the million dollar question follows from this and that’s about consciousness, qualia and all that jazz. Perhaps before we get into that, could you sketch out for us your approach to how the mind works? It’s kind of minimalist story you tell isn’t it?

AC:Well, not really minimalist. It’s true that ever since 1989 (Microcognition) I’ve embraced what I then called the 007 principle: know only what you need to know to get the job done. But my current story involves the bio-brain being a wonderfully canny player in the brain-body-world dance. In particular, I argue (in Surfing Uncertainty) that we benefit from lazy brains that use rich worlds models to estimate when a cheaper (embodied, extended) strategy is up to the current task.

3:AM:This has affinities with connectionism doesn’t it and the approach of the Churchlands? It goes up against the representational theorists like Fodordoesn’t it – is it because current AI is suggesting that we need to have far less representational capacity built in to the mind that sways you here?

AC:See previous answer. Expanding on that, I think we use deep world models to estimate when shallow ones – ones that may lean heavily on bio-external resources (thus happily recapturing the whole extended mind story) will do the job.

3:AM:How much of this model built on empirical evidence coming out of current cognitive science, psychology and AI etc? The science is really important isn’t it?

AC:Yes, good philosophy must be responsive to the best science of the day. I take that to be obvious.

3:AM:Doesn’t the Searlean Chinese Room argument and Leibniz’s Mill push back against the thought that really all we need are a complex system of cameras and other recording devices hooked up in a certain way to something like a computer for there to be consciousness? They suggest that there wouldn’t be consciousness no matter how many cameras, computers etc we had. Qualia would not appear. Now Dan Dennett’sposition argues that qualia are illusions brought on by the cameras and recording devices and computers.. The Churchlands seem to think that a theory about the structure of the brain will exhaust all that needs to be explained about consciousness and the qualia, if not illusions, will have to look after themselves (not sure I’ve got that right but you can correct me!). Dave Chalmers says Dennett is just ignoring the first person data and Paul Churchland’s position invites dualism. Where do you stand in this? Are you an illusionist or a dualist or somethingelse?

AC:Cameras don't understand. But if a camera had to match its current pixel level stimulation by re- creating it ‘from the top-down’ using stored knowledge about the world, that’s a whole new ball game. That, in a nutshell, is how I think understanding happens – see e.g. my ‘brains blog’ posts here.

That said, I do think many puzzles about consciousness and qualia are based on flawed pictures of what such ‘things’ might be.

3:AM:Dave Chalmers makes the point that dualism is a philosophical thesis and not a scientific theory and one that opens up new avenues of thinking that could eventually lead to a scientific research programme. Is this how you see one of the important ways science and philosophy work together?

AC:At a certain level yes. But our best current science offers no real solace that I can see for dualisms of any kind. Dave contemplates a dualism between information and matter, and that’s an intriguing case – but it all gets puzzling if, like me, you also think that information is only effective when implemented in a merely physical system, following merely physical laws.

3:AM:It seems like you’re working at providing us with a new paradigm of the mind, consciousness and eventually the self. Is that how you see what is happening at the moment and if so, how important is it that we do so?

AC:In the Clark and Criticsvolume, my lovely friend Jesse Prinz asks, in effect, what holds my various views together? What provides the common ground between my work on connectionism, robotics, embodied and extended cognition, and (most recently) ‘predictive processing’? His answer moved me. What Jesse suggests is that each new framework is really offering a new way of seeing ourselves, making everything spin, morph, and alter, around some new central construct, be it connectionism, cognitive extension, embodiment, or prediction. That could be worrying. How could it be good to be so fickle? But Jesse serves up a lovely and comforting thought, that each perspective is simply a kind of invitation to try out a way of seeing ourselves, and the place of mind in the universe. Perhaps no such picture can really hope to deliver the whole truth. But by fully inhabiting each picture, we just might learn something about who and what we are, and what we can hope to be. The exercise, Prinz suggests, is thus as much one of art as science. It is about bringing some stuff to the fore, and asking the reader to look at the world (even if only temporarily) through that lens.

3:AM:The idea of the singularity seems to be pressing in on us where machines overtake humans in some form another – Terminator scenarios, prosthetic minds where we become super-brained and the problem of consciousness is solved by four year olds and we live for thousands of years etc. As a philosopher leading this work do you think we should be thinking hard about the ethical issues that spring up from the new paradigms coming along? Are our ethical theories lagging behind the developments you’re working on, and are you optimistic or pessimistic about the way we’re going?

AC:I think we are playing ethical catch-up – but then we always have. I think I’m really just a techno- and human- optimist. As a global community, we are trying pretty hard to get better, and the tech can help as well as hinder.

3:AM:And finally, for those of us here at 3:AM, are there five books you can recommend we read to take us further into your philosophical world?

AC:From me:

Being There

and then Surfing Uncertainty.

From others: Susan Hurley’s Consciousness in Action,

Mike Wheeler’s, Reconstructing the Cognitive World,

and Michael Madary’s Visual Phenomenology.

Plus in the near future, look out for something ground-breaking from Anil Seth.

Richard Marshallis still biding his time.

Buy his new book hereor his first book hereto keep him biding!

End Times Series: the first 302