Diane Proudfoot

In the 70 years since Alan Turing’s ‘Computing Machinery and Intelligence’ appeared in Mind, there have been two widely-accepted interpretations of the Turing test: the canonical behaviourist interpretation and the rival inductive or epistemic interpretation. These readings are based on Turing’s Mind paper; few seem aware that Turing described two other versions of the imitation game. I have argued that both readings are inconsistent with Turing’s 1948 and 1952 statements about intelligence, and fail to explain the design of his game. I argue instead for a response-dependence interpretation. This interpretation has implications for Turing’s view of free will: I argue that Turing’s writings suggest a new form of free will compatibilism, which I call response-dependence compatibilism. The philosophical implications of rethinking Turing’s test go yet further. It is assumed by numerous theorists that Turing anticipated the computational theory of mind. On the contrary, I argue, his remarks on intelligence and free will lead to a new objection to computationalism.

Recent work in social robotics, which is aimed both at creating an artificial intelligence and providing a test-bed for psychological theories of human social development, involves building robots that can learn from ‘face-to-face’ interaction with human beings — as human infants do. The building-blocks of this interaction include the robot’s ‘expressive’ behaviours, for example, facial-expression and head-and-neck gesture. There is here an ideal opportunity to apply Wittgensteinian conceptual analysis to current theoretical and empirical work in the sciences. Wittgenstein’s philosophical psychology is sympathetic to embodied and situated Artificial Intelligence (see Proudfoot, 2002, 2004b), and his discussion of facial-expression is remarkably modern. In this chapter, I explore his approach to facial-expression, using smiling as a representative example, and apply it to the canonical interactive face robot, Cynthia Breazeal’s Kismet (see e.g. Breazeal, 2009, 2002). I assess the claim that Kismet has expressive behaviours, with the aim of generating philosophical insights for AI.

It is not widely realised that Turing was probably the first person to consider building computing machines out of simple, neuron-like elements connected together into networks in a largely random manner. Turing called his networks 'unorganised machines'. By the application of what he described as 'appropriate interference, mimicking education' an unorganised machine can be trained to perform any task that a Turing machine can carry out, provided the number of 'neurons' is sufficient. Turing proposed simulating both the behaviour of the network and the training process by means of a computer program. We outline Turing's connectionist project of 1948

Ignoring the temporal dimension, an object such as a railway tunnel or a human body is a three-dimensional whole composed of three-dimensional parts. The four-dimensionalist holds that a physical object exhibiting identity across time—Descartes, for example—is a four-dimensional whole composed of 'briefer' four-dimensional objects, its temporal parts. Peter van Inwagen (1990) has argued that four-dimensionalism cannot be sustained, or at best can be sustained only by a counterpart theorist. We argue that different schemes of individuation of temporal parts are available, which undermines van Inwagen's argument.

  Given (1) Wittgensteins externalist analysis of the distinction between following a rule and behaving in accordance with a rule, (2) prima facie connections between rule-following and psychological capacities, and (3) pragmatic issues about training, it follows that most, even all, future artificially intelligent computers and robots will not use language, possess concepts, or reason. This argument suggests that AIs traditional aim of building machines with minds, exemplified in current work on cognitive robotics, is in need of substantial revision

Alan Turing anticipated many areas of current research incomputer and cognitive science. This article outlines his contributionsto Artificial Intelligence, connectionism, hypercomputation, andArtificial Life, and also describes Turing's pioneering role in thedevelopment of electronic stored-program digital computers. It locatesthe origins of Artificial Intelligence in postwar Britain. It examinesthe intellectual connections between the work of Turing and ofWittgenstein in respect of their views on cognition, on machineintelligence, and on the relation between provability and truth. Wecriticise widespread and influential misunderstandings of theChurch–Turing thesis and of the halting theorem. We also explore theidea of hypercomputation, outlining a number of notional machines thatcompute the uncomputable.

Turing’s analysis of computability has recently been challenged; it is claimed that it is circular to analyse the intuitive concept of numerical computability in terms of the Turing machine. This claim threatens the view, canonical in mathematics and cognitive science, that the concept of a systematic procedure or algorithm is to be explicated by reference to the capacities of Turing machines. We defend Turing’s analysis against the challenge of ‘deviant encodings’.Keywords: Systematic procedure; Turing machine; Church–Turing thesis; Deviant encoding; Acceptable encoding; Turing’s analysis of computability; Turing’s Notational Thesis.

Cognitive science is held, not only by its practitioners, to offer something distinctively new in the philosophy of mind. This novelty is seen as the product of two factors. First, philosophy of mind takes itself to have well and truly jettisoned the ‘old paradigm’, the theory of the mind as embodied soul, easily and completely known through introspection but not amenable to scientific inquiry. This is replaced by the ‘new paradigm’, the theory of mind as neurally-instantiated computational mechanism, relatively opaque to introspection and the proper subject of detailed empirical investigation. Second, in the constitutive disciplines of cognitive science we have for the first time the theoretical, experimental and technological resources to begin this investigation. My concern here is to show that, despite its scientific and philosophical sophistication, the new paradigm is in certain striking ways very similar to the old paradigm and that Wittgenstein's criticisms of the former apply to much of the latter.

The widespread tendency, even within AI, to anthropomorphize machines makes it easier to convince us of their intelligence. How can any putative demonstration of intelligence in machines be trusted if the AI researcher readily succumbs to make-believe? This is (what I shall call) the forensic problem of anthropomorphism. I argue that the Turing test provides a solution. This paper illustrates the phenomenon of misplaced anthropomorphism and presents a new perspective on Turingʼs imitation game. It also examines the role of the Turing test in relation to the current dispute between human-level AI and ‘mindless intelligence’.

In this paper I argue that Turing proposed a new approach to the concept of thinking, based on his claim that intelligence is an ‘emotional concept’; and that the response-dependence interpretation of Turing’s ‘criterion for “thinking”’ is a better fit with his writings than orthodox interpretations. The aim of this paper is to clarify the response-dependence interpretation, by addressing such questions as: What did Turing mean by the expression ‘emotional’? Is Turing’s criterion subjective? Are ‘emotional’ judgements decided by social consensus? Turing’s take on these issues impacts current philosophical debates on response-dependent concepts and on the nature of artificial intelligence.

According to Richard Routley, a comprehensive theory of fiction is impossible, since almost anything is in principle imaginable. In my view, Routley is right: for any purported logic of fiction, there will be actual or imaginable fictions that successfully counterexample the logic. Using the example of ‘impossible’ fictions, I test this claim against theories proposed by Routley’s Meinongian contemporaries and also by Routley himself and his 21st century heirs. I argue that the phenomenon of impossible fictions challenges even today’s modal Meinongians.

Given actual autonomous systems with capacities for harm and the public’s apparent willingness to take moral advice from large language models (LLMs), Einar Duenger Bohn’s (2024) renewed discussion of the Moral Turing Test (MTT) is timely. Bohn’s aim is to defend an unequivocally behavioural test. In this paper, I argue against this direction. Interpreted as testing mere behaviour, the Turing test is a poor test of either intelligence or moral agency, and neither Bohn’s version of the test nor Allen, Varner and Zinser’s influential (2000) version avoids these problems. Also, the MTT’s advantages as advertised by Bohn do not hold up: it is an open empirical question whether embracing a merely behavioural MTT will significantly reduce the challenges in building a computer to pass the test. This issue aside, I argue that Turing’s actual test is superior to current AI benchmarks and to Bohn’s version of the test. A test of moral reasoning or agency in machines that is based on Turing’s actual test has advantages over current tests of moral reasoning in AI, including versions of the MTT. Such a test is an intriguing possibility still to be investigated.

Given actual autonomous systems with capacities for harm and the public’s apparent willingness to take moral advice from large language models (LLMs), Einar Duenger Bohn’s (2024) renewed discussion of the Moral Turing Test (MTT) is timely. Bohn’s aim is to defend an unequivocally behavioural test. In this paper, I argue against this direction. Interpreted as testing mere behaviour, the Turing test is a poor test of either intelligence or moral agency, and neither Bohn’s version of the test nor Allen, Varner and Zinser’s influential (2000) version avoids these problems. Also, the MTT’s advantages as advertised by Bohn do not hold up: it is an open empirical question whether embracing a merely behavioural MTT will significantly reduce the challenges in building a computer to pass the test. This issue aside, I argue that Turing’s actual test is superior to current AI benchmarks and to Bohn’s version of the test. A test of moral reasoning or agency in machines that is based on Turing’s actual test has advantages over current tests of moral reasoning in AI, including versions of the MTT. Such a test is an intriguing possibility still to be investigated.

The modern project of naturalizing intelligence began in the middle of last century, and Alan Turing is one of its most celebrated proponents. The assumption that Turing shared the ontological and methodological commitments of canonical naturalists is based on certain widespread beliefs about Turing—namely, that his test of intelligence is behaviourist and his approach to the mind computationalist. This chapter argues that influential versions of these assumptions are false, and instead that, in his claim that intelligence is an ‘emotional concept’, Turing proposed a response-dependence approach to intelligence. Recent objections, however, take the response-dependence interpretation to commit Turing to anti-realism—a philosophical stance that is typically anathema to naturalists. This chapter argues that those objections fail.

In 1948 Turing claimed that the concept of intelligence is an “emotional concept”. An emotional concept is a response-dependent concept and Turing’s remarks in his 1948 and 1952 papers suggest a response-dependence approach to the concept of intelligence. On this view, whether or not an object is intelligent is determined, as Turing said, “as much by our own state of mind and training as by the properties of the object”. His discussion of free will suggests a similar approach. Turing said, for example, that if a machine’s program “results in its doing something interesting which we had not anticipated I should be inclined to say that the machine had originated something”. This points to a new form of free will compatibilism, which I call response-dependence compatibilism and explore here.