This chapter summarizes the results that have been presented in the preceding chapters, in particular the process by which inference-driven mapping goes directly from the syntactic parse of a sentence to a “deep” semantic representation that corresponds to a traditional linguistic decomposition. The summary illustrates two of the most important benefits offered by inference-driven mapping over the template approach, namely, (1) the clear distinction between the verb definition and the final semantic representation achieved, and (2) an integrated approach to semantic analysis. The first benefit is of special relevance to linguistic theories about semantic representations, in that it provides a testing ground for such theories. The second benefit is of more relevance to computational models of natural language processors, in terms of interfacing semantic processing with syntactic parsing. The last section suggests directions of future research for pursuing these objectives.

Integrated semantic analysis

As discussed in chapter 2, traditional approaches to semantic processing need several levels of description to produce a “deep” semantic representation from a syntactic parse. The most popular of these approaches, termed the template approach, can be seen as using at least two intermediate levels of description, (1) the template level which is used for assigning mappings from syntactic constituents and semantic roles, and (2) the canonical level where the semantic roles are grouped together to simplify derivation of a “deep” semantic representation. These separate levels of description impose several stages of processing on the implementations, since only certain pieces of information are available at any one stage.

Introduction

Futurologists have proclaimed the birth of a new species, machina sapiens, that will share (perhaps usurp) our place as the intelligent sovereigns of our earthly domain. These “thinking machines” will take over our burdensome mental chores, just as their mechanical predecessors were intended to eliminate physical drudgery. Eventually they will apply their “ultra-intelligence” to solving all of our problems. Any thoughts of resisting this inevitable evolution is just a form of “speciesism,” born from a romantic and irrational attachment to the peculiarities of the human organism.

Critics have argued with equal fervor that “thinking machine” is an oxymoron – a contradiction in terms. Computers, with their foundations of cold logic, can never be creative or insightful or possess real judgment. No matter how competent they appear, they do not have the genuine intentionality that is at the heart of human understanding. The vain pretensions of those who seek to understand mind as computation can be dismissed as yet another demonstration of the arrogance of modern science.

Although my own understanding developed through active participation in artificial intelligence research, I have now come to recognize a larger grain of truth in the criticisms than in the enthusiastic predictions. But the story is more complex. The issues need not (perhaps cannot) be debated as fundamental questions concerning the place of humanity in the universe. Indeed, artificial intelligence has not achieved creativity, insight, and judgment. But its shortcomings are far more mundane: we have not yet been able to construct a machine with even a modicum of common sense or one that can converse on everyday topics in ordinary language.

Systems of interconnected and interdependent computers are qualitatively different from the relatively isolated computers of the past. Such “open systems” uncover important limitations in current approaches to artificial intelligence (AI). They require a new approach that is more like organizational designs and management than current approaches. Here we'll take a look at some of the implications and constraints imposed by open systems.

Open systems are always subject to communications and constraints from outside. They are characterized by the following properties:

Continuous change and evolution. Distributed systems are always adding new computers, users and software. As a result, systems must be able to change as the components and demands placed upon them change. Moreover, they must be able to evolve new internal components in order to accommodate the shifting work they perform. Without this capability, every system must reach the point where it can no longer expand to accommodate new users and uses.

Arm's-length relationships and decentralized decision making. In general, the computers, people, and agencies that make up open systems do not have direct access to one another's internal information. Arm's-length relationships imply that the architecture must accommodate multiple computers at different physical sites that do not have access to the internal components of others. This leads to decentralized decision making.

Perpetual inconsistency among knowledge bases. Because of privacy and discretionary concerns, different knowledge bases will contain different perspectives and conflicting beliefs. Thus, all the knowledge bases of a distributed AI system taken together will be perpetually inconsistent. Decentralization makes it impossible to update all knowledge bases simultaneously.

“But why,” Aunty asked with perceptible asperity, “does it have to be a language?” Aunty speaks with the voice of the Establishment, and her intransigence is something awful. She is, however, prepared to make certain concessions in the present case. First, she concedes that there are beliefs and desires and that there is a matter of fact about their intentional contents; there's a matter of fact, that is to say, about which proposition the intentional object of a belief or a desire is. Second, Aunty accepts the coherence of physicalism. It may be that believing and desiring will prove to be states of the brain, and if they do that's OK with Aunty. Third, she is prepared to concede that beliefs and desires have causal roles, and that overt behavior is typically the effect of complex interactions among these mental causes. (That Aunty was raised as a strict behaviorist goes without saying. But she hasn't been quite the same since the sixties. Which of us has?) In short, Aunty recognizes that psychological explanations need to postulate a network of causally related intentional states. “But why,” she asks with perceptible asperity, “does it have to be a language?” Or, to put it more succinctly than Aunty often does, what – over and above mere Intentional Realism – does the Language of Thought Hypothesis buy? That is what this discussion is about.

A prior question: what – over and above mere Intentional Realism – does the Language of Thought Hypothesis (LOT) claim? Here, I think, the situation is reasonably clear.

Introduction

Artificial intelligence is still a relatively young science, in which there are still various influences from different parent disciplines (psychology, philosophy, computer science, etc.). One symptom of this situation is the lack of any clearly defined way of carrying out research in the field (see D. McDermott, 1981, for some pertinent comments on this topic). There used to be a tendency for workers (particularly Ph.D. students) to indulge in what McCarthy has called the “look-ma-no-hands” approach (Hayes, 1975b), in which the worker writes a large, complex program, produces one or two impressive printouts and then writes papers stating that he has done this. The deficiency of this style of “research” is that it is theoretically sterile – it does not develop principles and does not clarify or define the real research problems. What has happened over recent years is that some attempt is now made to outline the principles which a program is supposed to implement. That is, the worker still constructs a complex program with impressive behaviour, but he also provides a statement of how it achieves this performance. Unfortunately, in some cases, the written “theory” may not correspond to the program in detail, but the writer avoids emphasizing (or sometimes even conceals) this discrepancy, resulting in methodological confusion. The “theory” is supposedly justified, or given empirical credibility, by the presence of the program (although the program may have been designed in a totally different way); hence the theory is not subjected to other forms of argument or examination.

Rational reconstruction (reproducing the essence of the program's significant behavior with another program constructed from descriptions of the purportedly important aspects of the original program) has been one approach to assessing the value of published claims about programs.

Campbell attempts to account for why the status of AI vis-a-vis the conventional sciences is a problematic issue. He outlines three classes of theories, the distinguishing elements of which are: equations; entities, operations and a set of axioms; and general principles capable of particularization in different forms. Models in AI, he claims, tend to fall in the last class of theory.

He argues for the methodology of rational reconstruction as an important component of a science of AI, even though the few attempts so far have not been particularly successful, if success is measured in terms of the similarity of behavior between the original AI system and the subsequent rational reconstruction. But, as Campbell points out, it is analysis and exploration of exactly these discrepancies that is likely to lead to significant progress in AI.

The second paper in this section is a reprint of one of the more celebrated attempts to analyse a famous AI program. In addition, to an analysis of the published descriptions of how the program works with respect to the program's behaviour (Lenat's ‘creative rediscovery’ system AM), Richie and Hanna discuss more general considerations of the rational-reconstruction methodology.

There is a continuing concern in AI that proof and correctness, the touchstones of the theory of programming, are being abandoned to the detriment of AI as a whole. On the other hand, we can find arguments to support just the opposite view, that attempts to fit AI programming into the specify-and-prove (or at least, specify-and-test correctness) paradigm of conventional software engineering, is contrary to the role of programming in AI research.

Similarly, the move to establish conventional logic as the foundational calculus of AI (currently seen in the logic programming approach and in knowledge-based decision-making implemented as a proof procedure) is another aspect of correctness in AI; and one whose validity is questioned (for example, Chandrasekaran's paper in section 1 opened the general discussion of such issues when it examined logic-based theories in AI, and Hewitt, in the section 11, takes up the more specific question of the role of logic in expert systems). Both sides of this correctness question are presented below.

Introduction

Philosophers constantly debate the nature of their discipline. These interminable debates frustrate even the most patient observer. Workers in AI also disagree, although not so frequently, about how to conduct their research. To equate programs with theories may offer a simple unifying tool to achieve agreement about the proper AI methodology. To construct a program becomes a way to construct a theory. When AI researchers need to justify their product as scientific, they can simply point to their successful programs. Unfortunately, methodological agreement does not come so easily in AI.

For a number of reasons, theorists in any discipline do not relish washing their proverbially dirty laundry in public. Methodology creates a great deal of that dirt, and philosophy of science supposedly supplies the soap to cleanse the discipline's methodology. Scientists often appeal to philosophers of science to develop methodological canons. It certainly does not instill confidence in a discipline if its practitioners cannot even agree on how to evaluate each other's work. Despite public images to the contrary, disagreements over how to approach a subject matter predominate in most scientific disciplines. Can philosophy of science come to the rescue of AI methodology? Yes and no.

Before describing the middle ground occupied by philosophy of science in its relationship to AI, we need to examine how some dismiss the AI research project altogether. In a previous article I argued against the various philosophical obstacles to the program/theory equation (Simon, 1979). I considered three types of objections to AI research: impossibility, ethical, and implausibility.

Introduction

The area of non-monotonic reasoning and the area of logic programming are of crucial and growing significance to artificial intelligence and to the whole field of computer science. It is therefore important to achieve a better understanding of the relationship existing between these two fields.

The major goal in the area of non-monotonic reasoning is to find adequate and sufficiently powerful formalizations of various types of non-monotonic reasoning – including common-sense reasoning – and to develop efficient ways of their implementation. Most of the currently existing formalizations are based on mathematical logic.

Logic programming introduced to computer science the important concept of declarative – as opposed to procedural – programming, based on mathematical logic. Logic programs, however, do not use logical negation, but instead rely on a non-monotonic operator – often referred to as negation as failure – which represents a procedural form of negation.

Non-monotonic reasoning and logic programming are closely related. The importance of logic programming to the area of non-monotontic reasoning follows from the fact that, as observed by several researchers (see e.g. Reiter, [to appear]) the non-monotonic character of procedural negation used in logic programming often makes it possible to efficiently implement other non-monotonic formalisms in Prolog or in other logic programming languages. Logic programming can also be used to provide formalizations for special forms of non-monotonic reasoning. For example, Kowalski and Sergot's calculus of events (1986) uses Prolog's negation-asfailure operator to formalize the temporal persistence problem in AI.

Fodor restates his language of thought hypothesis, which presents a serious challenge to the view that the architecture of cognition is network-based. Fodor claims that, whatever the representation underlying thought at the level of semantic interpretation, it must have constituent structure, i.e. a structure such that belief in the truth of (A and B) necessarily involves belief in the truth of both of the constituents, A and B, separately. Such necessity is not, in general, supported by network architecture whereas it is in Turing machine-type representations.

The rest of the papers in this section respond (directly or indirectly) to this challenge to the significance of connectionism in AI. Is connectionism just implementation detail, an essentially Turing machine-type architecture implemented with an activity-passing network? Or are sub-symbolic networks inherently more powerful than the traditional symbolic-level processing representations that have dominated much of AI? Smolensky is firmly on the side of connectionism as a fundamentally new and powerful ‘sub-symbolic’ paradigm for AI. Wilks discusses and denies the central claims of both Fodor and Smolensky, arguing that, at the moment, benificent agnosticism is the best position on connectionism, awaiting further clarification of its claims and more empirical results.

Churchland supports the connectionist movement but his support is based on the similarities between connectionist principles and the architecture of the brain. He argues for the study of neuroanatomy as a source of system-building constraints in AI.

Prolegomenon

My concern is with what an AI experiment is, and hence with what AI is. I shall talk about what experiments are actually like, but suggest that this is what they must be like.

Thus is it reasonable to suppose that AI experiments are, or could be, like the experiments of classical physics? I do not believe it is. This is not because we cannot expect the result of a single critical experiment to validate a theory, as we cannot expect a single translation to validate a translation program, for example: we can presumably extend the classical model to cover the case where validation depends on a set of results, for different data. Nor is it because we have not in practice got anything like an adequate predictive theory. I believe that we cannot in principle have the sort of predictive theory associated with physics, because we are not modelling nature in the classical physics sense. I shall elaborate on what I think we are doing, but claim now that we reach the same conclusion if we consider the suggestion that we are not in the classical physics position, but rather in that of investigative biologists, doing experiments to find out what nature is like (notionally without any theory at all, though perhaps in fact influenced by some half-baked theory). This is because there is nothing natural to discover.

Artificial intelligence is a subject that, due to the massive, often quite unintelligible, publicity that it gets, is nearly completely misunderstood by people outside the field. Even AIs practitioners are somewhat confused with respect to what AI is really about.

Is AI mathematics? A great many AI researchers believe strongly that knowledge representations used in AI programs must conform to previously established formalisms and logics or else the field will be unprincipled and ad hoc. Many AI researchers believe that they know how the answer will turn out before they have figured out what exactly the questions are. They know that some mathematical formalism or other must be the best way to express the contents of the knowledge that people have. Thus, to them, AI is an exercise in the search for the proper formalisms to use in representing knowledge.

Is AI software engineering? A great many AI practitioners seem to think so. If you can put knowledge into a program, then that program must be an AI program. This conception of AI, derived as it is from much of the work going on in industry in expert systems, has served to confuse AI people tremendously about what the correct focus of AI ought to be, and about what the fundamental issues in AI are.

A psychologist of my acquaintance, neither unaware of nor unsympathetic to work influenced by AI, recently referred to computational psychology as ‘very specialized’. Is this a fair assessment? Is Al-based psychology a mere hidden backwater, separated from the psychological mainstream? Sociologically, it must be admitted that it is. But theoretically? Perhaps the backwater is where the action is (and so much the worse for those who cannot see it)? Could it become the mainstream itself? In short, has AI helped psychology, or has it failed to live up to its early promise?

AI has undoubtedly helped psychology in some, very general, ways. It has provided a standard of rigour and completeness to which theoretical explanations should aspire (which is not to say that a program is in itself a theory). It has highlighted the importance of asking detailed questions about mental processes – about not just what the mind does, but how. It has alerted theoretical psychologists to the enormous richness and computational power of the mind. And it has furthered our understanding of how psychological, intentional, phenomena are possible in a basically material universe. If only for these reasons, AI has already been of lasting benefit to psychology. Certain sorts of theoretical inadequacies should be less common in the future than in the past.

From the beginning of computing people have been trying to make computers do more and more difficult things. At first these things were difficult in the sense that people could do them but rather tediously and unreliably, like solving lots of simultaneous equations. Then they were things that for practical purposes people could not do at all, like the calculations of theoretical chemistry. Progress was made by:

Researchers were slightly surprised to find that, when they wanted computers to do things that people find relatively easy, progress was much more difficult. This should perhaps be rephrased a bit. After all, people do not pick up objects in a three-pronged steel gripper, and they do not translate from French to English starting with the text punched on cards. Nor do they see through television cameras or hear through microphones. It was hard to progress with problems that looked as if they were in all material respects like things that people can do fairly easily.

Goldsmith's rustics were quite right about the village schoolmaster, of course, well in advance of their time and, apparently, of Goldsmith. But perhaps the time has come for less of such gazing, by AI researchers in particular, and more attention to their proper business. I am sure, for reasons I shall try to make clear, that the present situation, where much new work in AI is immediately proposed as new ‘model of the human brain or behaviour’, is an undesirable one.

This is a philosophical discussion about AI rather than a practical one; and intended to remind some AI researchers of standard uses of the words ‘model’ and ‘theory’ that they may have forgotten, and to explore some of the terminological consequences of a less liberal approach to these words than is the current fashion within AI.

Since it is a philosophical discussion, it is not intended to criticize any form of activity, or to suggest that it should not be carried out. It is concerned with how such work should be described, in the sense of what its ultimate subject matter is, and that such descriptions should be both revealing and consistent, and above all not misleading. I take it for granted in what follows that such questions are not “mere matters of words” or convention, and that it is no defence at any point to say that one can use the words ‘theory’ and ‘model’ to mean whatever one chooses.