Online research in philosophy

Turing''s test has been much misunderstood. Recently unpublished material by Turing casts fresh light on his thinking and dispels a number of philosophical myths concerning the Turing test. Properly understood, the Turing test withstands objections that are popularly believed to be fatal.

The Turing Test (TT) is claimed by many to be a way to test for the presence, in computers, of such ``deep'' phenomena as thought and consciousness. Unfortunately, attempts to build computational systems able to pass TT (or at least restricted versions of this test) have devolved into shallow symbol manipulation designed to, by hook or by crook, trick. The human creators of such systems know all too well that they have merely tried (...) to fool those people who interact with their systems into believing that these systems really have minds. And the problem is fundamental: the structure of the TT is such as to cultivate tricksters. A better test is one that insists on a certain restrictive epistemic relation between an artificial agent (or system) A, its output o, and the human architect H of A – a relation which, roughly speaking, obtains when H cannot account for how A produced o. We call this test the ``Lovelace Test'' in honor of Lady Lovelace, who believed that only when computers originate things should they be believed to have minds. (shrink)

The standard interpretation of the imitation game is defended over the rival gender interpretation though it is noted that Turing himself proposed several variations of his imitation game. The Turing test is then justified as an inductive test not as an operational definition as commonly suggested. Turing's famous prediction about his test being passed at the 70% level is disconfirmed by the results of the Loebner 2000 contest and the absence of (...) any serious Turing test competitors from AI on the horizon. But, reports of the death of the Turing test and AI are premature. AI continues to flourish and the test continues to play an important philosophical role in AI. Intelligence attribution, methodological, and visionary arguments are given in defense of a continuing role for the Turing test. With regard to Turing's predictions one is disconfirmed, one is confirmed, but another is still outstanding. (shrink)

I advocate a theory of syntactic semantics as a way of understanding how computers can think (and how the Chinese-Room-Argument objection to the Turing Test can be overcome): (1) Semantics, considered as the study of relations between symbols and meanings, can be turned into syntax – a study of relations among symbols (including meanings) – and hence syntax (i.e., symbol manipulation) can suffice for the semantical enterprise (contra Searle). (2) Semantics, considered as the process of understanding one domain (...) (by modeling it) in terms of another, can be viewed recursively: The base case of semantic understanding –understanding a domain in terms of itself – is syntactic understanding. (3) An internal (or narrow ), first-person point of view makes an external (or wide ), third-person point of view otiose for purposes of understanding cognition. (shrink)

The Turing Test is one of the most disputed topics in artificial intelligence, philosophy of mind, and cognitive science. This paper is a review of the past 50 years of the Turing Test. Philosophical debates, practical developments and repercussions in related disciplines are all covered. We discuss Turing''s ideas in detail and present the important comments that have been made on them. Within this context, behaviorism, consciousness, the `other minds'' problem, and similar topics (...) in philosophy of mind are discussed. We also cover the sociological and psychological aspects of the Turing Test. Finally, we look at the current situation and analyze programs that have been developed with the aim of passing the Turing Test. We conclude that the Turing Test has been, and will continue to be, an influential and controversial topic. (shrink)

The Turing Test is one of the most disputed topics in artificial intelligence, philosophy of mind, and cognitive science. This paper is a review of the past 50 years of the Turing Test. Philo- sophical debates, practical developments and repercussions in related disciplines are all covered. We discuss Turing’s ideas in detail and present the important comments that have been made on them. Within this context, behaviorism, consciousness, the ‘other minds’ problem, and similar (...) topics in philosophy of mind are discussed. We also cover the sociological and psychological aspects of the Turing Test. Finally, we look at the current situation and analyze programs that have been developed with the aim of passing the Turing Test. We conclude that the Turing Test has been, and will continue to be, an influential and controversial topic. (shrink)

The main factor of intelligence is defined as the ability tocomprehend, formalising this ability with the help of new constructsbased on descriptional complexity. The result is a comprehension test,or C-test, which is exclusively defined in computational terms. Due toits absolute and non-anthropomorphic character, it is equally applicableto both humans and non-humans. Moreover, it correlates with classicalpsychometric tests, thus establishing the first firm connection betweeninformation theoretical notions and traditional IQ tests. The TuringTest is compared with the C- (...) class='Hi'>test and the combination of the two isquestioned. In consequence, the idea of using the Turing Test as apractical test of intelligence should be surpassed, and substituted bycomputational and factorial tests of different cognitive abilities, amuch more useful approach for artificial intelligence progress and formany other intriguing questions that present themselves beyond theTuring Test. (shrink)

The Turing Test (TT), as originally specified, centres on theability to perform a social role. The TT can be seen as a test of anability to enter into normal human social dynamics. In this light itseems unlikely that such an entity can be wholly designed in anoff-line mode; rather a considerable period of training insitu would be required. The argument that since we can pass the TT,and our cognitive processes might be implemented as a (...) class='Hi'>Turing Machine(TM), that consequently a TM that could pass the TT could be built, isattacked on the grounds that not all TMs are constructible in a plannedway. This observation points towards the importance of developmentalprocesses that use random elements (e.g., evolution), but in these casesit becomes problematic to call the result artificial. This hasimplications for the means by which intelligent agents could bedeveloped. (shrink)

Stuart M. Shieber’s name is well known to computational linguists for his research and to computer scientists more generally for his debate on the Loebner Turing Test competition, which appeared a decade earlier in Communications of the ACM (Shieber 1994a, 1994b; Loebner 1994).¹ With this collection, I expect it to become equally well known to philosophers.

This paper argues that the Turing test is based on a fixed and de-contextualized view of communicative competence. According to this view, a machine that passes the test will be able to communicate effectively in a variety of other situations. But the de-contextualized view ignores the relationship between language and social context, or, to put it another way, the extent to which speakers respond dynamically to variations in discourse function, formality level, social distance/solidarity among participants, (...) and participants' relative degrees of power and status (Holmes, 1992). In the case of the Loebner Contest, a present day version of the Turing test, the social context of interaction can be interpreted in conflicting ways. For example, Loebner discourse is defined 1) as a friendly, casual conversation between two strangers of equal power, and 2) as a one-way transaction in which judges control the conversational floor in an attempt to expose contestants that are not human. This conflict in discourse function is irrelevant so long as the goal of the contest is to ensure that only thinking, human entities pass the test. But if the function of Loebner discourse is to encourage the production of software that can pass for human on the level of conversational ability, then the contest designers need to resolve this ambiguity in discourse function, and thus also come to terms with the kind of competence they are trying to measure. (shrink)

Alan Turing devised his famous test (TT) through a slight modificationof the parlor game in which a judge tries to ascertain the gender of twopeople who are only linguistically accessible. Stevan Harnad hasintroduced the Total TT, in which the judge can look at thecontestants in an attempt to determine which is a robot and which aperson. But what if we confront the judge with an animal, and arobot striving to pass for one, and then challenge him to peg (...) which iswhich? Now we can index TTT to a particular animal and its syntheticcorrelate. We might therefore have TTTrat, TTTcat,TTTdog, and so on. These tests, as we explain herein, are abetter barometer of artificial intelligence (AI) than Turing's originalTT, because AI seems to have ammunition sufficient only to reach thelevel of artificial animal, not artificial person. (shrink)

The test Turing proposed for machine intelligence is usually understood to be a test of whether a computer can fool a human into thinking that the computer is a human. This standard interpretation is rejected in favor of a test based on the Imitation Game introduced by Turing at the beginning of "Computing Machinery and Intelligence.".

Some of the papers in this special issue distribute cognition between what is going on inside individual cognizers' heads and their outside worlds; others distribute cognition among different individual cognizers. Turing's criterion for cognition was individual, autonomous input/output capacity. It is not clear that distributed cognition could pass the Turing Test.

The paper examines the nature of the behavioral evidence underlying attributions of intelligence in the case of human beings, and how this might be extended to other kinds of cognitive system, in the spirit of the original Turing Test (TT). I consider Harnad's Total Turing Test (TTT), which involves successful performance of both linguistic and robotic behavior, and which is often thought to incorporate the very same range of empirical data that is available in (...) the human case. However, I argue that the TTT is still too weak, because it only tests the capabilities of particular tokens within a preexisting context of intelligent behavior. What is needed is a test of the cognitive type, as manifested through a number of exemplary tokens, in order to confirm that the cognitive type is able to produce the context of intelligent behavior presupposed by tests such as the TT and TTT. (shrink)

The Turing Test is a verbal-behavioral operational criterion of artificial intelligence. If a machine can participate in question–and–answer conversation adequately enough to deceive an intelligent interlocutor, then it has intelligent information processing abilities. Robert M. French has argued that recent discoveries in cognitive science about subcognitive processes involving associational primings prove that the Turing Test cannot provide a satisfactory criterion of machine intelligence, that Turing's prediction concerning the feasibility of building machines to play the imitation (...) game successfully is false, and that the test should be rejected as ethnocentric and incapable of measuring kinds and degrees of nonhuman intelligence. But French's criticism is flawed, because it requires Turing's sufficient conditional criterion of intelligence to serve as a necessary condition. Turing's Test is defended against these objections, and French's claim that the test ought to be rejected because machines cannot pass it is deemed unscientific, resting on the empirically unwarranted assumption that intelligent machines are possible. (shrink)

This is my personal homage to Turing in his centenary anniversary. I don't deal with details of the Turing-Wittgenstein debate during the lectures on the foundation of Mathematics in '39, but I hint at a possible redefinition of Turing test inside a vision of thinking as use of symbols, in a (not new) Wittgensteinian fashion.

The Turing Test, originally proposed as a simple operational definition of intelligence, has now been with us for exactly half a century. It is safe to say that no other single article in computer science, and few other articles in science in general, have generated so much discussion. The present article chronicles the comments and controversy surrounding Turing's classic article from its publication to the present. The changing perception of the Turing Test over (...) the last fifty years has paralleled the changing attitudes in the scientific community towards artificial intelligence: from the unbridled optimism of 1960's to the current realization of the immense difficulties that still lie ahead. I conclude with the prediction that the Turing Test will remain important, not only as a landmark in the history of the development of intelligent machines, but also with real relevance to future generations of people living in a world in which the cognitive capacities of machines will be vastly greater than they are now. (shrink)

I have discussed the frame problem and the Turing test at length, but I have not attempted to spell out what I think the implications of the frame problem ...

The so-called Turing test, as it is usually interpreted, sets a benchmark standard for determining when we might call a machine intelligent. We can call a machine intelligent if the following is satisfied: if a group of wise observers were conversing with a machine through an exchange of typed messages, those observers could not tell whether they were talking to a human being or to a machine. To pass the test, the machine has to be intelligent but (...) it also should be responsive in a manner which cannot be distinguished from a human being. This standard interpretation presents the Turing test as a criterion for demarcating intelligent from non-intelligent entities. For a long time proponents of artificial intelligence have taken the Turing test as a goalpost for measuring progress. (shrink)

This commentary attempts to show that the inverted Turing Test (Watt 1996) could be simulated by a standard Turing test and, most importantly, claims that a very simple program with no intelligence whatsoever could be written that would pass the inverted Turing test. For this reason, the inverted Turing test in its present form must be rejected.

No computer that had not experienced the world as we humans had could pass a rigorously administered standard Turing Test. We show that the use of “subcognitive” questions allows the standard Turing Test to indirectly probe the human subcognitive associative concept network built up over a lifetime of experience with the world. Not only can this probing reveal differences in cognitive abilities, but crucially, even differences in _physical aspects_ of the candidates can be detected. (...) Consequently, it is unnecessary to propose even harder versions of the Test in which all physical and behavioral aspects of the two candidates had to be indistinguishable before allowing the machine to pass the Test. Any machine that passed the “simpler” symbols- in/symbols-out test as originally proposed by Turing would be intelligent. The problem is that, even in its original form, the Turing Test is already too hard and too anthropocentric for any machine that was not a physical, social, and behavioral carbon copy of ourselves to actually pass it. Consequently, the Turing Test, even in its standard version, is not a reasonable test for general machine intelligence. There is no need for an even stronger version of the Test. (shrink)

It is important to understand that the Turing Test (TT) is not, nor was it intended to be, a trick; how well one can fool someone is not a measure of scientific progress. The TT is an empirical criterion: It sets AI's empirical goal to be to generate human-scale performance capacity. This goal will be met when the candidate's performance is totally indistinguishable from a human's. Until then, the TT simply represents what it is that AI must (...) endeavor eventually to accomplish scientifically. (shrink)

After proposing the Turing Test, Alan Turing himself considered a number of objections to the idea that a machine might eventually pass it. One of the objections discussed by Turing was that no machine will ever pass the Turing Test because no machine will ever “have as much diversity of behaviour as a man”. He responded as follows: the “criticism that a machine cannot have much diversity of behaviour is just a way of saying (...) that it cannot have much storage capacity”. I shall argue that the objection cannot be dismissed so easily. The diversity exhibited by human behaviour is characterized by a kind of context-sensitive adaptive plasticity. Most of the time, human beings flexibly and fluently respond to what is relevant in a given situation. Moreover, ordinary human life involves an open-ended flow of shifting contexts to which our behaviour typically adapts in real time. For a machine to “have as much diversity of behaviour as a man” would be for that machine to keep its responses and behaviour relevant within such a flow. Merely giving a machine the capacity to store a huge amount of information and an enormous number of behaviour-generating rules will not achieve this goal. By drawing on arguments presented originally by Descartes, and by making contact with the frame problem in artificial intelligence, I shall argue that the distinctive context-sensitive adaptive plasticity of human behaviour explains why the Turing Test is such a stringent test for the presence of thought, and why it is much harder to pass than Turing himself may have realized. (shrink)

The paper begins by examining the original Turing Test (2T) and Searle’s antithetical Chinese Room Argument, which is intended to refute the 2T in particular, as well as any formal or abstract procedural theory of the mind in general. In the ensuing dispute between Searle and his own critics, I argue that Searle’s ‘internalist’ strategy is unable to deflect Dennett’s combined robotic-systems reply and the allied Total Turing Test (3T). Many would hold that the 3T marks (...) the culmination of the dialectic and, in principle, constitutes a fully adequate empirical standard for judging that an artifact is intelligent on a par with human beings. However, the paper carries the debate forward by arguing that the sociolinguistic factors highlighted in externalist views in the philosophy of language indicate the need for a fundamental shift in perspective in a Truly Total Turing Test (4T). It’s not enough to focus on Dennett’s individual robot viewed as a system; instead, we need to focus on an ongoing system of such artifacts. Hence a 4T should evaluate the general category of cognitive organization under investigation, rather than the performance of single specimens. From this comprehensive standpoint, the question is not whether an individual instance could simulate intelligent behavior within the context of a pre-existing sociolinguistic culture developed by the human cognitive type. Instead the key issue is whether the artificial cognitive type itself is capable of producing a comparable sociolinguistic medium. (shrink)

Stuart M. Shieber’s name is well known to computational linguists for his research and to computer scientists more generally for his debate on the Loebner Turing Test competition, which appeared a decade earlier in Communications of the ACM (Shieber 1994a, 1994b; Loebner 1994).¹ With this collection, I expect it to become equally well known to philosophers.

The Turing Test (TT) is criticised for various reasons, one being that it is limited to testing only human-like intelligence. We can read, for example, that ‘TT is testing humanity, not intelligence,’ (Fostel, 1993), that TT is ‘a test for human intelligence, not intelligence in general,’ (French, 1990), or that a perspective assumed by TT is parochial, arrogant and, generally, ‘massively anthropocentric’ (Hayes and Ford, 1996). This limitation presumably causes a basic inadequacy of TT, namely that it (...) misses a wide range of intelligence by focusing on one possibility only, namely on human intelligence. The spirit of TT enforces making explanations of possible machine intelligence in terms of what is known about intelligence in humans, thus possible specificity of the computer intelligence is ruled out from the oælset. (shrink)

The Turing Test blurs the distinction between a model and irrelevant) instantiation details. Modeling only functional modules is problematic if these are interconnected and cognitively penetrable.

Heyes's literature review of deception, imitation, and self-recognition is inadequate, misleading, and erroneous. The anaesthetic artifact hypothesis of self-recognition is unsupported by the data she herself examines. Her proposed experiment is tantalizing, indicating that theory of mind is simply a Turing test.

What would it take for an artificial agent to be treated as having moral value? As a first step toward answering this question, we ask what it would take for an artificial agent to be capable of the sort of autonomous, adaptive social behavior that is characteristic of the animals that humans interact with. We propose that this sort of capacity is best measured by what we call the Embodied Turing Test. The Embodied Turing test is (...) a test in which intelligence is operationally defined in terms of autonomous, adaptive interaction with the environment and with other animals. Three versions of the Embodied Turing test were performed with a SONY AIBO robot. Human participants were asked to differentiate between AIBO in a human-controlled mode and AIBO in a software-controlled mode. Our results indicate that the human participants were guessing at how AIBO was controlled. Our data reveals that people do not have enough experience with robots to accurately evaluate its behavior. This indicates that today’s humans do not have enough experience with artificial agents to treat them as morally valuable. (shrink)

This target article argues that the Turing test implicitly rests on a "naive psychology," a naturally evolved psychological faculty which is used to predict and understand the behaviour of others in complex societies. This natural faculty is an important and implicit bias in the observer's tendency to ascribe mentality to the system in the test. The paper analyses the effects of this naive psychology on the Turing test, both from the side of (...) the system and the side of the observer, and then proposes and justifies an inverted version of the test which allows the processes of ascription to be analysed more directly than in the standard version. (shrink)

In 1950, Alan Turing proposed his eponymous test based on indistinguishability of verbal behavior as a replacement for the question "Can machines think?" Since then, two mutually contradictory but well-founded attitudes towards the Turing Test have arisen in the philosophical literature. On the one hand is the attitude that has become philosophical conventional wisdom, viz., that the Turing Test is hopelessly flawed as a sufficient condition for intelligence, while on the other hand is the (...) overwhelming sense that were a machine to pass a real live full-fledged Turing Test, it would be a sign of nothing but our orneriness to deny it the attribution of intelligence. The arguments against the sufficiency of the Turing Test for determining intelligence rely on showing that some extra conditions are logically necessary for intelligence beyond the behavioral properties exhibited by an agent under a Turing Test. Therefore, it cannot follow logically from passing a Turing Test that the agent is intelligent. I argue that these extra conditions can be revealed by the Turing Test, so long as we allow a very slight weakening of the criterion from one of logical proof to one of statistical proof under weak realizability assumptions. The argument depends on the notion of interactive proof developed in theoretical computer science, along with some simple physical facts that constrain the information capacity of agents. Crucially, the weakening is so slight as to make no conceivable difference from a practical standpoint. Thus, the Gordian knot between the two opposing views of the sufficiency of the Turing Test can be cut. (shrink)

Robert French has argued that a disembodied computer is incapable of passing a Turing Test that includes subcognitive questions. Subcognitive questions are designed to probe the network of cultural and perceptual associations that humans naturally develop as we live, embodied and embedded in the world. In this paper, I show how it is possible for a disembodied computer to answer subcognitive questions appropriately, contrary to French’s claim. My approach to answering subcognitive questions is to use statistical information (...) extracted from a very large collection of text. In particular, I show how it is possible to answer a sample of subcognitive questions taken from French, by issuing queries to a search engine that indexes about 350 million Web pages. This simple algorithm may shed light on the nature of human (sub-) cognition, but the scope of this paper is limited to demonstrating that French is mistaken: a disembodied computer can answer subcognitive questions. (shrink)

The Turing Test is one of the most disputed topics in artificial intelligence, philosophy of mind, and cognitive science. This paper is a review of the past 50 years of the Turing Test. Philosophical debates, practical developments and repercussions in related disciplines are all covered. We discuss Turing's ideas in detail and present the important comments that have been made on them. Within this context, behaviorism, consciousness, the `other minds' problem, and similar topics in philosophy (...) of mind are discussed. We also cover the sociological and psychological aspects of the Turing Test. Finally, we look at the current situation and analyze programs that have been developed with the aim of passing the Turing Test. We conclude that the Turing Test has been, and will continue to be, an influential and controversial topic. (shrink)

I advocate a theory of syntactic semantics as a way of understanding how computers can think (and how the Chinese-Room-Argument objection to the Turing Test can be overcome): (1) Semantics, considered as the study of relations between symbols and meanings, can be turned into syntax – a study of relations among symbols (including meanings) – and hence syntax (i.e., symbol manipulation) can suffice for the semantical enterprise (contra Searle). (2) Semantics, considered as the process of understanding one (...) domain (by modeling it) in terms of another, can be viewed recursively: The base case of semantic understanding –understanding a domain in terms of itself – is syntactic understanding. (3) An internal (or narrow), first-person point of view makes an external (or wide), third-person point of view otiose for purposes of understanding cognition. (shrink)

Response to Floridi et al, 2008/2009. Based on insufficient evidence, and inadequate research, Floridi and his students report inaccuracies and draw false conclusions in their Minds and Machines evaluation, which this paper aims to clarify. Acting as invited judges, Floridi et al. participated in nine, of the ninety-six, Turing tests staged in the finals of the 18th Loebner Prize for Artificial Intelligence in October 2008. From the transcripts it appears that they used power over solidarity as an interrogation technique. (...) As a result, they were fooled on several occasions into believing that a machine was a human and that a human was a machine. Worse still, they did not realise their mistake. This resulted in a combined correct identification rate of less than 56%. In their paper they assumed that they had made correct identifications when they in fact had been incorrect. (shrink)

On a literal reading of `Computing Machinery and Intelligence'', Alan Turing presented not one, but two, practical tests to replace the question `Can machines think?'' He presented them as equivalent. I show here that the first test described in that much-discussed paper is in fact not equivalent to the second one, which has since become known as `the Turing Test''. The two tests can yield different results; it is the first, neglected test that provides the (...) more appropriate indication of intelligence. This is because the features of intelligence upon which it relies are resourcefulness and a critical attitude to one''s habitual responses; thus the test''s applicablity is not restricted to any particular species, nor does it presume any particular capacities. This is more appropriate because the question under consideration is what would count as machine intelligence. The first test realizes a possibility that philosophers have overlooked: a test that uses a human''s linguistic performance in setting an empirical test of intelligence, but does not make behavioral similarity to that performance the criterion of intelligence. Consequently, the first test is immune to many of the philosophical criticisms on the basis of which the (so-called) `Turing Test'' has been dismissed. (shrink)

Alan Turing devised his famous test (TT) through a slight modificationof the parlor game in which a judge tries to ascertain the gender of twopeople who are only linguistically accessible. Stevan Harnad hasintroduced the Total TT, in which the judge can look at thecontestants in an attempt to determine which is a robot and which aperson. But what if we confront the judge with an animal, and arobot striving to pass for one, and then challenge him to (...) peg which iswhich? Now we can index TTT to a particular animal and its syntheticcorrelate. We might therefore have TTTrat, TTTcat,TTTdog, and so on. These tests, as we explain herein, are abetter barometer of artificial intelligence (AI) than Turing's originalTT, because AI seems to have ammunition sufficient only to reach thelevel of artificial animal, not artificial person. (shrink)

The purpose of this paper is to consider Turing's two tests for machine intelligence: the parallel-paired, three-participants game presented in his 1950 paper, and the “jury-service” one-to-one measure described two years later in a radio broadcast. Both versions were instantiated in practical Turing tests during the 18th Loebner Prize for artificial intelligence hosted at the University of Reading, UK, in October 2008. This involved jury-service tests in the preliminary phase and parallel-paired in the final phase.

Turing's celebrated 1950 paper proposes a very general methodological criterion for modelling mental function: total functional equivalence and indistinguishability. His criterion gives rise to a hierarchy of Turing Tests, from subtotal ("toy") fragments of our functions (t1), to total symbolic (pen-pal) function (T2 -- the standard Turing Test), to total external sensorimotor (robotic) function (T3), to total internal microfunction (T4), to total indistinguishability in every empirically discernible respect (T5). This is a "reverse-engineering" hierarchy of (decreasing) empirical (...) underdetermination of the theory by the data. Level t1 is clearly too underdetermined, T2 is vulnerable to a counterexample (Searle's Chinese Room Argument), and T4 and T5 are arbitrarily overdetermined. Hence T3 is the appropriate target level for cognitive science. When it is reached, however, there will still remain more unanswerable questions than when Physics reaches its Grand Unified Theory of Everything (GUTE), because of the mind/body problem and the other-minds problem, both of which are inherent in this empirical domain, even though Turing hardly mentions them. (shrink)

If, as a number of writers have predicted, the computers of the future will possess intelligence and capacities that exceed our own then it seems as though they will be worthy of a moral respect at least equal to, and perhaps greater than, human beings. In this paper I propose a test to determine when we have reached that point. Inspired by Alan Turing’s (1950) original “Turing test”, which argued that we would be justified (...) in conceding that machines could think if they could fill the role of a person in a conversation, I propose a test for when computers have achieved moral standing by asking when a computer might take the place of a human being in a moral dilemma, such as a “triage” situation in which a choice must be made as to which of two human lives to save. We will know that machines have achieved moral standing comparable to a human when the replacement of one of these people with an artificial intelligence leaves the character of the dilemma intact. That is, when we might sometimes judge that it is reasonable to preserve the continuing existence of a machine over the life of a human being. This is the “Turing Triage Test”. I argue that if personhood is understood as a matter of possessing a set of important cognitive capacities then it seems likely that future AIs will be able to pass this test. However this conclusion serves as a reductio of this account of the nature of persons. I set out an alternative account of the nature of persons, which places the concept of a person at the centre of an interdependent network of moral and affective responses, such as remorse, grief and sympathy. I argue that according to this second, superior, account of the nature of persons, machines will be unable to pass the Turing Triage Test until they possess bodies and faces with expressive capacities akin to those of the human form. (shrink)

This paper presents an analysis of three major contests for machine intelligence. We conclude that a new era for Turing’s test requires a fillip in the guise of a committed sponsor, not unlike DARPA, funders of the successful 2007 Urban Challenge.

A series of imitation games involving 3-participant (simultaneous comparison of two hidden entities) and 2-participant (direct interrogation of a hidden entity) were conducted at Bletchley Park on the 100th anniversary of Alan Turing’s birth: 23 June 2012. From the ongoing analysis of over 150 games involving (expert and non-expert, males and females, adults and child) judges, machines and hidden humans (foils for the machines), we present six particular conversations that took place between human judges and a hidden entity that (...) produced unexpected results. From this sample we focus on features of Turing’s machine intelligence test that the mathematician/code breaker did not consider in his examination for machine thinking: the subjective nature of attributing intelligence to another mind. (shrink)

Searle's celebrated Chinese Room Argument has shaken the foundations of Artificial Intelligence. Many refutations have been attempted, but none seem convincing. This paper is an attempt to sort out explicitly the assumptions and the logical, methodological and empirical points of disagreement. Searle is shown to have underestimated some features of computer modeling, but the heart of the issue turns out to be an empirical question about the scope and limits of the purely symbolic (computational) model of the mind. Nonsymbolic modeling (...) turns out to be immune to the Chinese Room Argument. The issues discussed include the Total Turing Test, modularity, neural modeling, robotics, causality and the symbol-grounding problem. (shrink)

Both Artificial Life and Artificial Mind are branches of what Dennett has called "reverse engineering": Ordinary engineering attempts to build systems to meet certain functional specifications, reverse bioengineering attempts to understand how systems that have already been built by the Blind Watchmaker work. Computational modelling (virtual life) can capture the formal principles of life, perhaps predict and explain it completely, but it can no more be alive than a virtual forest fire can be hot. In itself, a computational model is (...) just an ungrounded symbol system; no matter how closely it matches the properties of what is being modelled, it matches them only formally, with the mediation of an interpretation. Synthetic life is not open to this objection, but it is still an open question how close a functional equivalence is needed in order to capture life. Close enough to fool the Blind Watchmaker is probably close enough, but would that require molecular indistinguishability, and if so, do we really need to go that far? (shrink)

We confront the following popular views: that mind or life are algorithms; that thinking, or more generally any process other than computation, is computation; that anything other than a working brain can have thoughts; that anything other than a biological organism can be alive; that form and function are independent of matter; that sufficiently accurate simulations are just as genuine as the real things they imitate; and that the Turing test is either a necessary or sufficient or (...) scientific procedure for evaluating whether or not an entity is intelligent. Drawing on the distinction between activities and tasks, and the fundamental scientific principles of ontological lawfulness, epistemological realism, and methodological skepticism, we argue for traditional scientific materialism of the emergentist kind in opposition to the functionalism, behaviourism, tacit idealism, and merely decorative materialism of the artificial intelligence and artificial life communities. (shrink)

I propose to consider the question, "Can machines think?" This should begin with definitions of the meaning of the terms "machine" and "think." The definitions might be framed so as to reflect so far as possible the normal use of the words, but this attitude is dangerous, If the meaning of the words "machine" and "think" are to be found by examining how they are commonly used it is difficult to escape the conclusion that the meaning and the answer to (...) the question, "Can machines think?" is to be sought in a statistical survey such as a Gallup poll. But this is absurd. Instead of attempting such a definition I shall replace the question by another, which is closely related to it and is expressed in relatively unambiguous words. ...". (shrink)

Many debates about consciousness appear to be endless, in part because of conceptual confusions preventing clarity as to what the issues are and what does or does not count as evidence. This makes it hard to decide what should go into a machine if it is to be described as 'conscious'. Thus, triumphant demonstrations by some AI developers may be regarded by others as proving nothing of interest because the system does not satisfy *their* definitions or requirements specifications.

This paper is a follow-up of the first part of the persons reply to the Chinese Room Argument. The first part claims that the mental properties of the person appearing in that argument are what matter to whether computational cognitive science is true. This paper tries to discern what those mental properties are by applying a series of hypothetical psychological and strengthened Turing tests to the person, and argues that the results support the thesis that the Man performing the (...) computations characteristic of understanding Chinese actually understands Chinese. The supposition that the Man does not understand Chinese has gone virtually unquestioned in this foundational debate. The persons reply acknowledges the intuitive power behind that supposition, but knows that brute intuitions are not epistemically sacrosanct. Like many intuitions humans have had, and later deposed, this intuition does not withstand experimental scrutiny. The second part of the persons reply consequently holds that computational cognitive science is confirmed by the Chinese Room thought experiment. (shrink)

An Evaluation of the 2008 Loebner Contest.

Brain damage can cause massive changes in consciousness levels. From a clinical and ethical point of view it is desirable to assess the level of residual consciousness in unresponsive patients. However, no direct measure of consciousness exists, so we run into the philosophical problem of other minds. Neurologists often make implicit use of a Turing test-like procedure in an attempt to gain access to damaged minds, by monitoring and interpreting neurobehavioral responses. New brain imaging techniques are now being (...) developed that permit communication with unresponsive patients, using their brain signals as carriers of messages relating to their mental states. (shrink)

We critically discuss Cleland''s analysis of effective procedures as mundane effective procedures. She argues that Turing machines cannot carry out mundane procedures, since Turing machines are abstract entities and therefore cannot generate the causal processes that are generated by mundane procedures. We argue that if Turing machines cannot enter the physical world, then it is hard to see how Cleland''s mundane procedures can enter the world of numbers. Hence her arguments against versions of the Church-Turing thesis (...) for number theoretic functions miss the mark. (shrink)

Alan Turing advocated a kind of functionalism: A machine M is a thinker provided that it responds in certain ways to certain inputs. Davidson argues that Turing’s functionalism is inconsistent with a certain kind of epistemic externalism, and is therefore false. In Davidson’s view, concepts consist of causal liasons of a certain kind between subject and object. Turing’s machine doesn’t have the right kinds of causal liasons to its environment. Therefore it doesn’t have concepts. Therefore it doesn’t (...) think. I argue that this reasoning is entirely fallacious. It is true that, in some cases, a causal liason between subject and object is part of one’s concept of that object. Consequently, to grasp certain propositions, one must have certain kids of causal ties to one’s environment. But this means that we must rethink some old views on what rationality is. It does not mean, pace Davidson, that a precondition for being rational is being causally embedded in one’s environment in a certain way. If Turing’s machine isn’t capable of thinking (I leave it open whether it is or is not), that has nothing to do with its lacking certain kinds of causal connections to the environment. The larger significance of our discussion is this: rationality consists either in one’s ability to see the bearing of purely existential propositions on one another or rationality is simply not to be understood as the ability see the bearing that propositions have on one another. (shrink)

In 1949, the Department of Philosophy at the University of Manchester organized a symposium “Mind and Machine” with Michael Polanyi, the mathematicians Alan Turing and Max Newman, the neurologists Geoff rey Jeff erson and J. Z. Young, and others as participants. Th is event is known among Turing scholars, because it laid the seed for Turing’s famous paper on “Computing Machinery and Intelligence”, but it is scarcely documented. Here, the transcript of this event, together with Polanyi’s original (...) statement and his notes taken at a lecture by Jeff erson, are edited and commented for the fi rst time. Th e originals are in the Regenstein Library of the University of Chicago. Th e introduction highlights elements of the debate that included neurophysiology, mathematics, the mind-body-machine problem, and consciousness and shows that Turing’s approach, as documented here, does not lend itself to reductionism. (shrink)

A fundamental problem in artificial intelligence is that nobody really knows what intelligence is. The problem is especially acute when we need to consider artificial systems which are significantly different to humans. In this paper we approach this problem in the following way: we take a number of well known informal definitions of human intelligence that have been given by experts, and extract their essential features. These are then mathematically formalised to produce a general measure of intelligence for arbitrary machines. (...) We believe that this equation formally captures the concept of machine intelligence in the broadest reasonable sense. We then show how this formal definition is related to the theory of universal optimal learning agents. Finally, we survey the many other tests and definitions of intelligence that have been proposed for machines. (shrink)

John Searle’s Chinese room argument (CRA) is a celebrated thought experiment designed to refute the hypothesis, popular among artificial intelligence (AI) scientists and philosophers of mind, that “the appropriately programmed computer really is a mind”. Since its publication in 1980, the CRA has evoked an enormous amount of debate about its implications for machine intelligence, the functionalist philosophy of mind, theories of consciousness, etc. Although the general consensus among commentators is that the CRA is flawed, and not withstanding the popularity (...) of the systems reply in some quarters, there is remarkably little agreement on exactly how and why it is flawed. A newcomer to the controversy could be forgiven for thinking that the bewildering collection of diverse replies to Searle betrays a tendency to unprincipled, ad hoc argumentation and, thereby, a weakness in the opposition’s case. In this paper, treating the CRA as a prototypical example of a ‘destructive’ thought experiment, I attempt to set it in a logical framework (due to Sorensen), which allows us to systematise and classify the various objections. Since thought experiments are always posed in narrative form, formal logic by itself cannot fully capture the controversy. On the contrary, much also hinges on how one translates between the informal everyday language in which the CRA was initially framed and formal logic and, in particular, on the specific conception(s) of possibility that one reads into the logical formalism. (shrink)

In this paper, I argue that certain Machines can have rights independently of whether they are sentient, or conscious, or whatever you might call it.

John Searle has argued that the aim of strong AI of creating a thinking computer is misguided. Searle’s Chinese Room Argument purports to show that syntax does not suffice for semantics and that computer programs as such must fail to have intrinsic intentionality. But we are not mainly interested in the program itself but rather the implementation of the program in some material. It does not follow by necessity from the fact that computer programs are defined syntactically that the implementation (...) of them cannot suffice for semantics. Perhaps our world is a world in which any implementation of the right computer program will create a system with intrinsic intentionality, in which case Searle’s Chinese Room Scenario is empirically (nomically) impossible. But, indeed, perhaps our world is a world in which Searle’s Chinese Room Scenario is empirically (nomically) possible and that the silicon basis of modern day computers are one kind of material unsuited to give you intrinsic intentionality. The metaphysical question turns out to be a question of what kind of world we are in and I argue that in this respect we do not know our modal address. The Modal Address Argument does not ensure that strong AI will succeed, but it shows that Searle’s challenge on the research program of strong AI fails in its objectives. (shrink)

A. M. Turing has bequeathed us a conceptulary including 'Turing, or Turing-Church, thesis', 'Turing machine', 'universal Turing machine', 'Turing test' and 'Turing structures', plus other unnamed achievements. These include a proof that any formal language adequate to express arithmetic contains undecidable formulas, as well as achievements in computer science, artificial intelligence, mathematics, biology, and cognitive science. Here it is argued that these achievements hang together and have prospered well in the 50 years (...) since Turing's death. (shrink)

As is well known, Alan Turing drew a line, embodied in the "Turing test," between intellectual and physical abilities, and hence between cognitive and natural sciences. Less familiarly, he proposed that one way to produce a "passer" would be to educate a "child machine," equating the experimenter's improvements in the initial structure of the child machine with genetic mutations, while supposing that the experimenter might achieve improvements more expeditiously than natural selection. On the other hand, in his (...) foundational "On the chemical basis of morphogenesis," Turing insisted that biological explanation clearly confine itself to purely physical and chemical means, eschewing vitalist and teleological talk entirely and hewing to D'Arcy Thompson's line that "evolutionary 'explanations,'" are historical and narrative in character, employing the same intentional and teleological vocabulary we use in doing human history, and hence, while perhaps on occasion of heuristic value, are not part of biology as a natural science. To apply Turing's program to recent issues, the attempt to give foundations to the social and cognitive sciences in the "real science" of evolutionary biology (as opposed to Turing's biology) is neither to give foundations, nor to achieve the unification of the social/cognitive sciences and the natural sciences. (shrink)

In the 1950s, Alan Turing proposed his influential test for machine intelligence, which involved a teletyped dialogue between a human player, a machine, and an interrogator. Two readings of Turing''s rules for the test have been given. According to the standard reading of Turing''s words, the goal of the interrogator was to discover which was the human being and which was the machine, while the goal of the machine was to be indistinguishable from a human (...) being. According to the literal reading, the goal of the machine was to simulate a man imitating a woman, while the interrogator – unaware of the real purpose of the test – was attempting to determine which of the two contestants was the woman and which was the man. The present work offers a study of Turing''s rules for the test in the context of his advocated purpose and his other texts. The conclusion is that there are several independent and mutually reinforcing lines of evidence that support the standard reading, while fitting the literal reading in Turing''s work faces severe interpretative difficulties. So, the controversy over Turing''s rules should be settled in favor of the standard reading. (shrink)

Computation is interpretable symbol manipulation. Symbols are objects that are manipulated on the basis of rules operating only on theirshapes, which are arbitrary in relation to what they can be interpreted as meaning. Even if one accepts the Church/Turing Thesis that computation is unique, universal and very near omnipotent, not everything is a computer, because not everything can be given a systematic interpretation; and certainly everything can''t be givenevery systematic interpretation. But even after computers and computation have been successfully (...) distinguished from other kinds of things, mental states will not just be the implementations of the right symbol systems, because of the symbol grounding problem: The interpretation of a symbol system is not intrinsic to the system; it is projected onto it by the interpreter. This is not true of our thoughts. We must accordingly be more than just computers. My guess is that the meanings of our symbols are grounded in the substrate of our robotic capacity to interact with that real world of objects, events and states of affairs that our symbols are systematically interpretable as being about. (shrink)

To clarify the notion of computation and its role in cognitive science, we need an account of implementation, the nexus between abstract computations and physical systems. I provide such an account, based on the idea that a physical system implements a computation if the causal structure of the system mirrors the formal structure of the computation. The account is developed for the class of combinatorial-state automata, but is sufficiently general to cover all other discrete computational formalisms. The implementation relation is (...) non-vacuous, so that criticisms by Searle and others fail. This account of computation can be extended to justify the foundational role of computation in artificial intelligence and cognitive science. (shrink)

This quote/commented critique of Turing's classical paper suggests that Turing meant -- or should have meant -- the robotic version of the Turing Test (and not just the email version). Moreover, any dynamic system (that we design and understand) can be a candidate, not just a computational one. Turing also dismisses the other-minds problem and the mind/body problem too quickly. They are at the heart of both the problem he is addressing and the solution he (...) is proposing. (shrink)