Science in Christian Perspective
Artificial Intelligence Research: An Evangelical Assessment
Terry A. Ward
Academic Computing Services
University of Northern Iowa
Cedar Falls, Iowa 50614
From: JASA 35 (March 1983): 39-42.
"What has Jerusalem to do with Athens, the Church with the Academy, the Christian with the Heretic?" (Tertullian)
The above quotation from Tertullian may be the first impression one has when considering the relationship between artificial intelligence, its concomitant philosophical underpinnings and the biblical, evangelical view of humanity. Artificial intelligence seems to be either an esoteric discipline of the computer scientist or the literary ramblings of the science-fiction author in a pulp novel. The biblical and evangelical perspective on man seems to belong more to the realm of theological speculation and theological science. The two seem to have little or no relation to the other.
It is my opinion that the relationship of these two fields highlights several areas of conflict between modern manifestations of the social sciences (as exemplified in artificial intelligence (AI) research and methodology) and the traditional, biblically based view of one's personhood.
It is my thesis that research in artificial intelligence represents the logical outcome of social science assumptions concerning humanity and the means used to understand our social existence. The impact of artificial intelligence upon our concepts of ourselves is sufficiently important that one writer has remarked, "literature on robotics is nothing less than a debate on the meaning and purpose of existence" (Hexharn 1980; p. 574). For S. L. Jaki (1969; p. 11), artificial intelligence represents the latest phase of scientism or physicalism, an approach that owes its impetus to the general success of physics and assigns virtually unlimited wisdom to the physical sciences.
I am also concerned with the extent to which artificial intelligence assumptions color our thinking. This is reflected in the popular literature of the day. These assumptions are rarely, if ever, addressed from a theological perspective.
An example of the popular (and widely held) view that computers and brains are identical is that expressed by Lees in Creative Computing (1977), "One way or another, we are going to run into another intelligence before much longer ... we will not remain alone." One such intelligence considered most likely by Lees is that of artificial machine intelligence. A recent book blithely assumes the ultimate equality of personality and mechanism in its very title, Machines Who Think (McCorduck 1979). In this text of the history of artificial intelligence research, we hear that Al is to be a "new species on the horizon" and that this development will be "very humbling to us." These sentiments represent a widely held view that any evangelical analysis must deal with. This view is not limited to the relatively small computer science or electrical engineering community. A very popular writer in the genre of science popularization, Carl Sagan, has argued that Al is inevitable and is simply an extension of evolutionist thought (Sagan 1977). Finally, a recent winner of the National Book Award, R. Hofstadter, simply dismisses those who doubt the possibility of artificial intelligence (in his book, Gbdel, Escher, Bach).
Lest you think my Protestations over the apparent religious fervor present in the writings of some authors is a misguided fear, I point to the work of Desmonde. According to Desmonde, ". . . computers will come to grips with all value problems . . . Mankind has to look to artificial intelligence for a new world where life will be meaningful" (Desmonde 1964; p. 5). This overblown rhetoric is not contained in an obscure (and hence relatively harmless) technical report or memorandum, but is rather presented in the introductory chapter of one of the most widely used textbooks in introductory computer science courses! The warning of Jaki thus seems to be particularly appropriate, ". . . the danger of machines derives not so much from the machines as from man's attitudes toward them" (Jaki 1969; p. 254). The use of the computer as metaphor for (or in many cases the exact equivalent to) the human mind and personality is also reflected in such works as The Human Machine (Aleksander 1978), The Brain as Computer (George 1961) or The Metaphorical Brain (Arbib 1972). In reference to works of this type, Jaki has remarked that "scientific fashions ... get hold of the minds of several generations with no small damage to cultural values" (Jaki 1969; p. 9). 1 tend to concur with this view as it applies to the currently fashionable reductionistic social sciences, particularly as this reductionism is manifested in the artificial intelligence literature. This essay is a small attempt at answering the often implicit and frequently anti-Christian assumptions underlying this line of research.
Before beginning our analysis of the theological perspective that evangelical Christianity can apply to the Al literature, it might be helpful to point out several works that are useful for an assessment a( the history of this computer science field. A detailed examination of the history of Al research can be found in either McCorduck (19791 or Dreyfus (1979). The former presents a quite sympathetic picture of the history and prospects for Al research, while Dreyfus presents an openly critical analysis. The basic thesis in Dreyfus, one with which I concur, is that the "field of artificial intelligence exhibits a recurring pattern: early dramatic successes followed by sudden unexpected difficulties" (Dreyfus 1979; p. 85). In addition, Dreyfus maintains (and, I feel, correctly) that there are innate man-machine differences in perception and that these qualitative differences render machine intelligence impossible. A similar thesis concerning the impossibility of artificial intelligence can be found in Weizenbaum (1976). Both Dreyfus and Weizenbaum, by the way, are professional computer scientists and were once both active in the Al research community. Finally, we also agree with Jaki that the literature of artificial intelligence blatantly overestimates computers and dramatically underestimates-man (Jaki 1969).
While making no claim to exhaustiveness or definitiveness, it is nonetheless hoped that these thoughts may contribute a theological perspective from which to view and assess the current world view envisaged by the "artificial intelligentsia" (to borrow a phrase from Weizenbaum).
"If he should take back his spirit to himself, and gather to himself his breath, all flesh would perish together, and man would return to dust." (Job 34:14-15)
The above question, from that most human book of the Bible, Job, highlights the dilemma -faced by anyone hoping to construct a distinctively Christian anthropology; namely, man is at one and the same time a creature made of dust, and yet, in the sight of God, a special creature in all of creation. This paradox or dilemma must always be kept in mind as we consider the theological and biblical issues and evidences.
Failure to maintain this paradox can lead either to a materialistic reductionism where, in the words of Arthur Koestler, the "anthropomorphic view of rats has been replaced by a ratomorphic view of man;" or to the converse danger in which the biblical insistence upons man's physical nature is replaced by a virtually docetic view of man (as in Descartes, for example).
One might be tempted to object that the relationship between cybernetics (or artificial intelligence research) and theology is a forced one; the two are separate and non-interacting fields of human endeavour. I reply only with that quotation from Hexharn with which I am in complete agreement, "literature on robotics is nothing less than a debate on the meaning and purpose of existence" (Hexham 1980).
The Bible contains numerous words and phrases that might be applicable to our consideration of the nature of man's psyche. The first that we consider is nous (reason, mind). In classical Greek thought and usage, reason represented the central area of thought. (Plato for example equates the process of thought with a calculating process.) Plato also used the term (Phaedr. 247c) to refer to the ruling principle of pure thought and of the divine reason that resides in man. In this usage, the term nous has come down to our Western culture as that phase of thought that is basically docetic in its view of the body. This "disembodiment" of intelligence from the human body is an essential assumption of the artificial intelligence community. It is also crucial to note that this conception derives, not from biblical considerations, but from classical Greek thought.
Niebuhr has considered this position and has termed it the classical view of man. According to Niebuhr, this view argues that, ". . . man is to be understood primarily from the standpoint of his rational faculties" (Niebuhr 1949; p. 6). In his view, this attitude is essentially dualistic and rationalistic. If we consider the history of artificial intelligence research (cf. McCorduck, Weizenbaum or Dreyfus), we can see that this opinion of Niebuhr concerning the classical view of man is essentially a correct assessment of the All community.
When we turn to the Scriptures themselves, we see that the word nous has taken on rather different meanings and connotations (J. Goetzmann, "Reason in TNIDNTT 1978 vol. 3; pp. 122-133).
According to Goetzmann, Old Testament anthropology knows nothing of the Greek division of the soul or psyche into three distinct elements (cf. p. 124). The Old Testament understanding of "reason" is that it belongs to the will, and that its aim is less at a theoretical comprehension than at right conduct.
Likewise, in the New Testament the concept of nous is not central. For example, in twenty-one of the twenty-four uses in the Pauline corpus, it means simply religious understanding (e.g. 2 Thes. 2:12). Similarly, in the pastorals, the term nous is used as a virtual equivalent to "religion" or "faith." Examples of this latter usage may be found in I Tim. 6:6 and 2 Tim. 3:8. Finally, in Rom. 11:34 the phrase nous theos is used as a phrase to mean God's plan of salvation itself. Thus, in considering the word nous, we see that it means an essentially religious understanding and not a calculating function. More germane to our analysis of the biblical evidence pertaining to the claims of artificial intelligence research is the following observation. The entire thrust of the biblical conception of humanity is of a unified individual; there is no disembodiment of soul (or mind) from the body as there is in the case of classical Greek thought or in the assumptions of artificial intellience research.
A second word that may be of interest in our investigation of biblical evidence is dialogismos (think, reason, etc.). The term is used extensively in the New Testament (e.g. Rom. 2:3; Gal. 3:6; 2 Tim. 4:16). For Paul, the principle function of this word is in relating the foundations of faith to the righteousness of God (cf. J. Eichler "Think" in TNIDNTT vol. 3; pp. 820-825). For Paul, this faith is not an objective reasoning (or in the terminology of computer science, "computable") from a neutral (or "context- free") vantage point. Rather, faith is the process of being conquered by the crucified and risen Lord. The discrete, unambiguous hopes embodied in the assumptions of the Al endeavors pale in the light of the "scandal" of the cross. We should always keep in mind Moltmann's admonition that "Christian anthropology [is always] an anthropology of the crucified Lord" (Moltmann 1974; p. 20). The discussion up to this point has revealed that the biblical concepts relevant to "mind" seem to have a much greater existential content than an analytical one, to be more related to the realm of faith than to the cold world of calculation.
The term phroneo, which is generally translated as "mind," is also a non-technical term used in the New Testament that pertains to the realm of faith. As Goetzmann argues (J. Goetzmann, "Mind" in TNIDNTT vol. 2; pp. 616-620), the word comes quite near that of the Old Testament concept of "wisdom," and "such wisdon ... is not merely rational discernment but is the result of God's people being turned to the obedience of faith, in accordance with Old'Testament prophecy."
Finally in this consideration of the linguistic evidence we come to the various words that are translated as "man" (cf. C. Brown "Man" in TNIDNTT, vol. 2; pp. 562-572). Here, as in the case of nous, we can see the influence of classical Greek (not biblical) thought on the modern world view as espoused in the artificial intelligence community.
The basic Greek conception of the body-soul duality is present in the Apocrypha (e.g. 4 Macc. 2:21f) as well as the extrabiblical literature of the early New Testament period. As an example, Josephus (War 2, 8) speaks of the image of God as being present in man only within his nous (in this usage, only within his rational facilities).
This is quite a contrast to the Old Testament's essential physicality and the New Testament's lack of abstractions concerning the human psyche. We are seen to be different from animals and plants (I Cor. 15:39); humans are also different from angels (I Cor. 4:9). Most importantly, we are different from Christ (Gal. 1: 12; Eph. 6:7) and from God (Matt. 7:11; John 10:33; Acts 5:29). On the whole, Brown concludes that "Man is seen as a whole being, and whatever touches one part affects the whole" (TNIDNTTvol. 2; p. 567f). This unitary view of man is noted by Ladd in his study of Paul when Ladd argues that "Paul never uses 'psyche' as a separate entity and he never speaks of the soul's survival of death" (Ladd 1974; pp. 457-459).
This essential unity of man's psyche and body is of fundamental importance in contrasting the artificial intelligence claims that man's essence is simply rationality and the rest of man is of greatly diminished importance. The biblical evidence will not allow this bifurcation of one's faculties as in the Al conceptions.
Having considered the strictly biblical evidences as shown by word usage, we now consider the various theological issues raised by claims of the artificial intelligence community.
We agree with Dreyfus (1979) and Weizenbaum (1976) that the claims of artificial intelligence researchers are greatly exaggerated. However, this incompatibility of claim vis-a-vis results is not without its danger to biblical perspectives. Scientific fashions (irrespective of their validity) have a way of capturing the imagination.
The first area of concern in the assumptions or theological implications of artificial intelligence research is in the area of man's unitary nature. The artificial intelligence research endeavor is essentially docetic in orientation, with its belief that man's essence is contained in and defined by his rationality (or more precisely in his "information-processing" capability). This position reflects a lengthy development in Western philosophical thought, ultimately deriving from classical Greek and not biblical perspectives.
Tillich, for example, traces this development as a process comprising two phases. First, he sees the development of what he calls "ontological reason." This is soon replaced in the western world-view with a concept of "technical reason." (Tillich 1957; vol. 1; pp. 70-75). It is this latter that is stressed within the artificial intelligence community. As defined by Tillich, technical reason is that tradition of philosophy which stresses only the cognitive side of reason (as an example he lists the English empiricists). He further adds that "Technical reason, however refined in logical and methodological aspects, dehumanizes man if it is separated from ontological reason" (Tillich 1957; vol. 1; p. 73). This emphasis upon the calculable or "technical reason," as is evident in the artificial intelligence research effort, is incompatible with the view of man as presented in the Scriptures. As Bloesch points out, this emphasis is a culmination of rationalism where, in the latter, ". . . the real is rational and truth signifies the whole" (Bloesch 1971; p. 69). This rationalism is in contrast with our insistence that faith precedes understanding (as in Augustine). As Bloesch phrases it, "Faith ... is an acquaintance with Christ rather than a comprehension of Him" (Bloesch 197 1; p. 68).
While we have disagreed with the conception of the artificial intelligence community in its assessment of mankind, which is that rationality is the "measure of man," we have not limited our conception of faith to the purely emotional or ecstatic. The realm of faith consists of both an objective and subjective pole. To stress the former can lead to a sterile scholasticism, while overemphasizing the latter can lead to a content-less enthusiasm. As Torrance phrases this conceptual basis of faith,
Knowledge of God is thus conceptual in its essential root with a conceptuality that derives from God's self revelation in Word, but which we bring to articulate expression in our understanding with a conceptuality that finds shape in our human forms of thought and speech yet under the control of God's own intelligible reality. (Torrance, 1971; pp. 21-22)
Faith always precedes understanding; ". . . the Christian life ... must be grounded in the inwardness of faith ... revelation can only be grasped by the inwardness of faith, and apart from this subjective dimension the objective data have little or no value." (Bloesch 1979, vol. 2; p. 259)
This is the fundamental difference between the artificial intelligence approach to man and a Christian anthropology. For the artificial intelligence community, man is essentially a rational calculating mechanism. For the Christian, man is a being created in the image of God and created to respond to his Creator.
This essential "respondability" of man is stressed by both systematic and biblical theologians. Berkhof agrees with this assessment of true human nature when he argues that the core of human nature is respondability (Berkhof 1979; p. 168). Ladd argues that, for Paul, man's true essence lies in the pneuma; "his inner self that is able to allow his relating to God" (Ladd 1974; p. 461-463).
According to Tillich, the essence of Christian faith is absolutely paradoxical (without the eyes of faith) and is hence completely foreign to the conception of man present in the assumptions undergirding Al research. He writes, "The Christian assertion that the New Being has appeared in Jesus the Christ is paradoxical . . ." (Tillich 1957; vol. 2; pp. 90-92). This, however, is the paradox upon which we base our entire existence!
Other differences between the contents of the Al image of man and a Christian anthropology could be outlined, but they pale in comparison with these two major and irreconcilable contrasts. The first of these two differences is the unitary view of humankind in the Christian anthropology (recall our unanimous confession of faith in the resurrection of the body), as contrasted with the rationalistic reductionism evident in the assumptions of artificial intelligence advocates. The second contrast occurs with the conception that, in the artificial intelligence view, rationality is the hope of our "salvation," (if I may use that phrase) whereas, for the Christian, faith and our personal relationship to our Lord and Savior precedes our limited understanding of God's plan and purpose for our life.
In closing, I would like simply to repeat an observation that Jaki has made concerning the nature of mankind that shows a great similarity to the biblical concern outlined above (Jaki 1969):
... man, precisely because of his mind, is not a machine, but a marvel and ought to be treated as such.
So God created man in his own image, in the image of God he created him; male and female he created them. And God blessed them ... (Gen 1:27-28a)
Aleksander, 1. (1978). The Human Machine. (Saphorin, Switzerland: Georgi Publishing Co. 1978).Arbib, M.A. (1972). The Metaphorical Brain. (New York: Wiley, 1972).
Berkbof, H. (1979). Christian Faith., trans. S. Woudstra. (Grand Rapids, MI: Eerdmans, 1979).
Bloesch, D.G. (1971). The Ground of Certainty. (Grand Rapids, MI: Eerdmans, 1971).
---. (1979). Essentials of Evangelical Theology., 2 vols. (San Francisco: Harper & Row, 1979).
Desmonde, W,H. (1964). Computers and Their Uses. (Englewood Cliffs, NJ. Prentice-Hall, 1964).
Dreyfus, H.L. (1979). What Computers Can't Do. (San Francisco: Harper & Row, 1979).George, F.H. (1961). The Brain as a Computer, (New York: Pergamon, 1961
Hexham, 1. (1980). "Learning to Live with Robots." Christian Century 97:574 (May 21, 1980).Hofstadter, D. R. (1979). Godel, Escher, Bach. (New York: Basic Books, 1979.
Jaki, S.L. (1969). Brain, Mind and Computers. (South Bend, IN: Gatewa% Editions, 1969).
Ladd, G.E. (1974). A Theology of the New Testament. (Grand Rapids Eerdmans, 1974).
Lees, John. (1977). "An Esoteric Ethical Excursion" and "A Return to Freedom and Dignity" in The Best of Creative Computing, vol. 2; D. Ahl, ed (Morristown: Creative Computing, 1979).
McCorduck, P. (1979). Machines Who Think. (San Francisco: W.H. FreemarL 1979).
Moltmann, J. (1974). Man., trans. J. Sturd. (Philadelphia: Fortress Press1974).Sagan, C. (1977). The Dragons of Eden. (New York: Random House, 1977).
Tillich, T. (1957). Systematic Theology. (Chicago: University of Chicago Press. 1957).
TNIDNTT . (1978). The New International Dictionary of New Testament Theology. C. Brown, ed. (Grand Rapids: Zondervan, 1978).Torrance, T.F. (1971). God and Rationality. (New York: Oxford, 1971).
Weizenbaurn, J. (1976). Computer Power and Human Reason. (San Francisco:
W.H. Freeman, 1976).