**Stephen Hawking at Google’s annual Zeitgeist**

On 15th May 2011, Stephen Hawking gave a fascinating talk at Google’s annual Zeitgeist.

Amongst some debatable and personally revealing and controversial remarks that Hawking made during the conference, was his extraordinarily provocative pronouncement—almost en passant—in the early moments of his talk:

“Philosophy is dead”!

I only came to know of Hawking’s talk—and of the continuing disturbed reactions to it from some philosophers—last week when I asked to join Professor Anthony Beaver’s LinkedIn group ‘Computing and Philosophy’.

Their discussion gave me considerable food for thought as I found myself—like I suspect quite a few of them—as uncomfortable dismissing Hawking’s remarks out of hand as I was in trying to place them in a perspective that did justice both to Hawking and to those exercised by his utterance, without being apologetic of any of their perspectives.

I was faced with the challenge:

**How do I honestly view Hawking’s remarks?**

So, to help me crystalise, and in the process share, my thoughts with any interested putative reader of this blog (since I don’t know if there actually is one at the moment), please allow me to make an arbitrary distinction between the following three disciplines (a distinction which I have personally found useful in the past when attempting to convey my thoughts to others, particularly to an inter-disciplinary audience):

**Applied science**, whose concern is our sensory observations of a ‘common’ external world;

**Philosophy**, whose concern is abstracting a coherent perspective of the external world from our sensory observations; and

**Mathematics**, whose concern is adequately expressing such abstractions in a formal language of unambiguous communication.

I would see these disciplines as broadly addressing the questions of:

**What** we do,

**Why** we do what we do, and

**How** we express and communicate whatever it is that we do for whatever reason.

**Was Hawking’s remark irresponsible?**

The reason for requesting indulgence in the making of such distinctions up front is that I have no reason to believe that any one else would naturally see them similarly; so it is perhaps worthwhile laying out the groundwork for my immediate observation (after listening to Stephen Hawking’s talk), which is that:

Hawking has not—as perhaps he responsibly should have—provided a suitable groundwork for helping place his unarguably philosophical comments in an appropriate context.

So perhaps given that—by the very nature of the occasion that invited his comments—Hawking *ought* to have been aware of his eminence and the influence of his comments on the scientific laity, he could—and possibly should, as some views in this Computing and Philosophy discussion (such as those of Jorge Baralt) appear to suggest—be held guilty of unnecessarily challenging his peers to reconstruct his specific intent from his general—and arguably loose—remarks.

If so, that is a guilt shared by all humanity.

That aside, it is difficult to conceive that Hawking’s remark—that philosophy is dead—was intended towards any discipline of ‘Philosophy’ that is consistent with what I have distinguished above.

**Was Hawking’s remark academically disrespectful?**

Of course, if Hawking’s remarks were indeed intended to implicitly censure the larger body of those whose primary interest lies in questioning and trying to gain a holistic perspective on *why* we do what we do, and *not* so much in the substance of what we believe we have achieved in any single direction, then perhaps Hawking could—and possibly should, as some views in this discussion (for instance those of Teed Rockwell and Patrick Dursi) appear to suggest—justly be accused of indulging in a moment of intellectual arrogance and academic disrespect.

Again, to be guilty of such indulgences is merely to be human.

**Was Hawking’s remark merely reflecting a subjective belief**

‘If’, because if Hawking’s implicit intent was to censure *only* the smaller body of philosophers who are focused on questioning why physicists and cosmologists do what they do, then perhaps we should not deny him his hard fought right to censure the perspective of his philosophical peers from his own, subjective, vision of what he perceives as his reasons for doing what he does.

Although we could—and perhaps should as some views in this discussion (for instance those of Robert Ware) appear to suggest—withhold from him the authority to pass *as* sweeping a generalisation on what their vision ought to be as he seems to implicitly pronounce.

**Was Hawking’s remark a refutation Lucas’ Gödelian argument?**

Or perhaps what Hawking intended to say—obliquely refuting Lucas’ Gödelian argument—was that the human mind must of necessity recognise that it will eventually depend upon a mechanical intelligence (read technology) as the final arbiter of all that we have so far treated as reliable, self-evident, truth.

If so, he is perhaps both a prisoner and promoter of a classical inheritance that, whilst acknowledging the limits on the capacity of human intelligence to predict quantum phenomena, has yet to acknowlege that human intelligence—which must of necessity reflect natural law—may also be an arbiter of an algorithmically verifiable truth that lies beyond the arbitration of the algorithmically computable truth that is, first, the realm of mechanical intelligences and, second, one by which both such intelligences and those who perhaps feel the need to depend upon them as the final arbiters of all scientific truth, are inexorably bound.

If so, it is a point—and counter point—whose complications I have attempted to address elsewhere.

**Or is Hawking, perhaps, simply tired of fence-sitters**

Or perhaps Hawking has just reached the age where he is perhaps simply tired of those who insist on prefacing each opinion with a ‘perhaps’; and perhaps he needs to seek comfort from those who are perhaps more willing to commit themselves to positions that, perhaps, give some more of a meaning to each man’s search for a meaning that he can, perhaps, give to his life’s work.

So perhaps—and again as some views in this discussion appear to reflect—we may need to keep in mind the dictum of Edmund Landau, who in 1929 sagely advised prospective seekers of truth in his ‘Foundations of Analysis’ that, in order to isolate such truth, one ought to consciously disbelieve all that one has been told; the truth must of necessity lie in the unconscious remains that resist the strongest of disbelief.

## 5 comments

Comments feed for this article

September 23, 2013 at 2:01 pm

Vivek IyerPerhaps you have come across Beenakker’s solution to Hempel’s dillemma- http://arxiv.org/abs/physics/0702072v1? Essentially, the notion is that Universe does not last long enough to solve some problems so they are metaphysical.

There are 3 reasons why Philosophy can’t die

1) Suppose we have arisen from natural selection. Then the question arises ‘how do we get enough epigenetic canalisation so as to be able to talk about Mathis or Physics or whatever.’ Now one way to solve a Schelling co-ordination problem is to look at the Muth rational solution- i.e. a canonical solution of a ‘philosophical’ type because it treats of something abstract philosophical term like ‘the virtuous thing to do’ or ‘the just thing to do’ or ‘the beautiful thing to do’. In other words the very fact we have Philosophy as a regulative discipline and moreover that it has co-evolved complexity with Physics suggests that it can’t die. I think it’s a parasite which militates towards a type of ‘Baldwin effect’ whereby smart people run away from a field where philosophers start pontificating- e.g. Bergson or Heidegger’s theory of Time or Amartya Sen’s type of Economics.

2) Suppose our Minds are only accidentally attached to bodies and the material world- i.e. suppose some Occassionalist type of set up. In this case Collingwood’s argument re. Philosophy’s ‘distinctions without a difference’ gain salience, indeed they become the only thing worth thinking about since it is immaterial how or why such distinctions can be instantiated. This argument holds even if we don’t adopt Mind/Body duality. We just say Philosophy’s ‘distinctions without a difference’ are part of capacitance diversity.

3) Information asymmetry- even if we were all computer simulations, still we can encode something personal to ourselves such that information asymmetry exists. This means there is a missing market for Philosophy to cater to- i.e. the demand will call forth a supply. True Philosophy will be competing with Astrology and Numerology and other such nonsense- but it will be able to make a modest living.

October 6, 2013 at 4:25 pm

Bhupinder Singh AnandDear Vivek,

An intriguing perspective!

I particularly like the idea of casting Philosophy into the role of the judiciary, presumably balancing the roles of Physics as the lawmaker, and of Mathematics as the executive that makes the rules for expressing and interpreting the laws.

It’s a great thesis for a future post.

October 20, 2013 at 9:47 am

Vivek IyerIt is interesting, in the light of what you have said, that the Just King, Yuddhishtra can only overcome his ‘vishada’ or depression by learning Mathematics- including Statistical Game Theory- thus fitting himself to rule. Arjuna, was more lucky. Since both he and Lord Krishna, appear in the Gita, merely as Agents, not as Principals, that sublime poem can be appreciated by ordinary people. We might say the duties of the Agent are ‘algorithmically verifiable but not computable’ (at least for non Laplacian intelligences) because we can just look them up in the appropriate Law book. However, as Lord Krishna later points out (when Arjun wants to kill Yuddhishtra) ‘Dharma is very complex. You (Arjuna) don’t understand it’. In fact, the mathematical structure of the Mahabharata- where every action is doubled, every agent has a dual- looks like a non-dissipative system governed by Noether’s theorem. Of course it has to be like that because the heuristic rules by which it was built up had to be simple and similar to double entry book-keeping. Thus both ‘karma’- i.e. the law of causation across time- and ‘dharma’- the same thing but across space- become conserved properties of the system by reason of the symmetrical manner in which the Epic was built up.

I know Andrei Weil was fascinated with the Mahabharata- he learned Sanskrit and came to India- but Indian game theorists have fought shy of looking at it in this light.

The reason I bring up all this ancient history is that it is suggestive that a crisis in ‘Sankhya’ philosophy (or was it some sort of discrete maths/O.R type bundle of Social heurisitcs?) leads in different directions to both Buddhism and Vaishnavite theism both of which say that there is an easily verifiable ‘middle’ or ‘golden path’ but its computability is beyond mortal scope.

Still, just as the anthropic principle may be able to give Scientist’s information about fundamental laws of the Universe- i.e. since we exist, some otherwise unobservable constant can be taken to hold- might it not be the case that some Mathematical truths, or at least Research Programs (like yours) could arise from purely anthropic arguments? True, this sort of ‘Kantian’ thinking has proven a stumbling block in the past- and my sympathy is with Hawking, especially in light of Hilary Putnam or Lee Smolin’s strange assertions- still,it seems there is some deep mystery in the connection between the naturals and the reals which, it may be, autistic savants, or perhaps even mystics, can glimpse.

On a different subject, if I may be permitted an ignorant query re. your latest post-

‘If we take both L and M as PA (as detailed in ‘Evidence-Based Interpretations of PA‘), and take satisfiability in PA to mean instantiational provability in PA, we arrive at the formal definition of the truth of the PA-formula [(\forall x)R(x)] in PA as:

The PA-formula [(\forall x)R(x)] is formally true in PA if, and only if, the formula [R(x)] is provable in PA whenever we substitute a numeral [n] for the variable [x] in [R(x)].

Surely this formula isn’t formally true in PA but rather in the M of its M?

Also,

‘for (vii), then, implicitly implies that the arithmetical relation R(x) is algorithmically decidable as always true in the standard interpretation of PA’

Why?

I have some hazy notion that the above would mean something strong about computably enumerable reals.

November 5, 2013 at 12:02 am

Bhupinder Singh Anand“

… it seems there is some deep mystery in the connection between the naturals and the reals which, it may be, autistic savants, or perhaps even mystics, can glimpse.”The focus of my investigation is not on the possible connection between what the naturals and the reals can be assumed to represent under an interpretation, but on how the connection between the formal definitions of the naturals and the reals limit what they can be assumed to represent under an interpretation.

The former is a question for philosophers to ponder upon, and physical scientists to uncover.

The latter is a purely mathematical one that determines the rules for any interpretation.

“

Surely this formula isn’t formally true in PA but rather in the M of its M?”Again, the focus is not upon whether or not a formula is true under an interpretation, but on defining a constructive and objective convention under which we could agree upon its ‘truth’ under an interpretation.

“

Why?”As to why (vii), then, implicitly implies that the arithmetical relation R(x) is algorithmically decidable as always true in the standard interpretation of PA, the answer lies in the standard formulation of the Church-Turing Thesis.

This would equate both algorithmic verifiabilty and algorithmic computability with the undefined intuitive concept of effective computability.

Now, if we

defineeffective computability as algorithmic verifiability, then CT would be false.“

I have some hazy notion that the above would mean something strong about computably enumerable reals.”In arithmetic, the computably enumerable reals are indeed the algorithmically verifiable reals.

In set theory, however, the distinction between computable (recursive) sets and computably (recursively) enumerable sets is not well-defined.

Reason: Since sets are defined extensionally in set theory, the theory cannot distinguish that two number-theoretic functions, such as below, could be computationally distinct, and may represent different situations under an interpretation.

For instance, by Theorem VII of Gödel’s famous 1931 paper on undecidable arithmetical propositions, it can be shown (see Lemmas 8 and 9 of this paper) that every algorithmically computable (recursive) number-theoretic function f(x) is instantiationally equivalent to an arithmetical function g(x) which is algorithmically verifiable but not algorithmically computable.

The distinction could be of significance in a game where Alice knows how to algorithmically compute f(x), but Bob only knows how to algorithmically verify g(x).

November 10, 2013 at 6:37 am

Vivek IyerMany thanks