Essays in Experimental Logic · John Dewey
SCIENCE AS A PRACTICAL ART
No one will deny that, as seen from one angle science is a pursuit, an enterprise--a mode of practice. It is at least that, no matter how much more or else it is. In course of the practice of knowing distinctive practical judgments will then naturally be made. Especially does this hold good when an intellectual class is developed, when there is a body of persons working at knowing as another body is working at farming or engineering. Moreover, the instrumentalities of this inquiring class gain in importance for all classes in the degree in which it is realized that success in the conduct of the practice of farming or engineering or medicine depends upon use of the successes achieved in the business of knowing. The importance of the latter is thrown into relief from another angle if we consider the enterprises, like diplomacy, politics, and, to a considerable extent, morals, which do not acknowledge a thoroughgoing and constant dependence upon the practice of science. As Hobbes was wont to say, the advantages of a science of morals are most obvious in the evils which we suffer from its lack.
To say that something is to be learned, is to be found out, is to be ascertained or proved or believed, is to say that something is to be done. Every such proposition in the concrete is a practical proposition. Every such proposition of inquiry, discovery and testing will have then the traits assigned to the class of practical propositions. They imply an incomplete situation going forward to completion, and the proposition as a specific organ of carrying on the movement. I have not the intention of dwelling at length upon this theme. I wish to raise in as definite and emphatic a way as possible a certain question. Suppose that the propositions arising within the _practice_ of knowing and functioning as agencies in its conduct could be shown to present all the distinctions and relations characteristic of the subject-matter of logic: what would be the conclusion? To an unbiased mind the question probably answers itself: All purely logical terms and propositions fall within the scope of the class of propositions of inquiry as a special form of propositions of practice. My further remarks are not aimed at _proving_ that the case accords with the hypothesis propounded, but are intended to procure hospitality for the hypothesis.
If thinking is the art by which knowledge is practiced, then the materials with which thinking deals may be supposed, by analogy with the other arts, to take on in consequence special shapes. The man who is making a boat will give wood a form which it did not have, in order that it may serve the purposes to which it is to be put. Thinking may then be supposed to give its material the form which will make it amenable to its purpose--attaining knowledge, or, as it is ordinarily put, going from the unknown to the known. That physical analysis and synthesis are included in the processes of investigation of natural objects makes them a part of the practice of knowing. And it makes any general traits which result in consequence of such treatment characters of _objects as they are involved in knowledge-getting_. That is to say, if there are any features which natural existences assume in order that inference may be more fertile and more safe than it would otherwise be, those features correspond to the special traits which would be given to wood in process of constructing a boat. They are manufactured, without being any worse because of it. The question which I raised in the last paragraph may then be restated in this fashion: Are there such features? If there are, are they like those characters which books on logic talk about?
Comparison with language may help us. Language--I confine myself for convenience to spoken language--consists of sounds. But it does not consist simply of those sounds which issue from the human organs prior to the attempt to communicate. It has been said that an American baby before talking makes almost every sound found in any language. But elimination takes place. And so does intensification. Certain sounds originally slurred over are made prominent; the baby has to work for them and the work is one which he neither undertakes nor accomplishes except under the incitation of others. Language is chiefly marked off, however, by articulation; by the arrangement of what is selected into an orderly sequence of vowels and consonants with certain rules of stress, etc. It may fairly be said that speech is a manufactured article: it consists of natural ebullitions of sound which have been shaped for the sake of being effective instrumentalities of a purpose. For the most part the making has gone on under the stress of the necessities of communication with little deliberate control. Works on phonetics, dictionaries, grammars, rhetorics, etc., mark some participation of deliberate intention in the process of manufacture. If we bring written language into the account, we should find the conscious factor extended somewhat. But making, shaping for an end, there is, whether with or without conscious control.
Now while there is something in the antecedent properties of sound which enters into the determination of speech, the _worth_ of speech is in no way measured by faithfulness to these antecedent properties. It is measured only by its efficiency and economy in realizing the special results for which it is constructed. Written language need not look like sounds any more than sounds look like objects. It must _represent_ articulate sounds, but faithful representation is wholly a matter of carrying the mind to the same outcome, of exercising the same function, not of resemblance or copying. Original structure _limits_ what may be made out of anything: one cannot (at least at present) make a silk purse out of pigs' bristles. But this conditioning relationship is very different from one in which the antecedent existences are a model or prototype to which the consequent must be servilely faithful. The boatmaker must take account of the grain and strength of his wood. To take account of, to reckon with, is a very different matter, however, from repetition or literal loyalty. The measure is found in the consequences for which existences are used.
I wish, of course, to suggest that logical traits are just features of original existences as they have been worked over for use in inference, as the traits of manufactured articles are qualities of crude materials modified for specific purposes. Upon the whole, past theories have vibrated between treating logical traits as "subjective," something resident in "mind" (mind being thought of as an immaterial or psychical existence independent of natural things and events), and ascribing ontological pre-existence to them. Thus far in the history of thought, each method has flourished awhile and then called out a reaction to its opposite. The reification (I use the word here without prejudice) of logical traits has taken both an Idealistic form (because of emphasis upon their spiritual or ideal nature and stuff) and a Realistic one, due to emphasis upon their immediate apprehension and givenness. That mathematics have been from Plato to Descartes and contemporary analytic realism the great provocative of Realistic Idealisms is a familiar fact. The hypothesis here propounded is a _via media_. What has been overlooked is the reality and importance of art and its works. The tools and works of art are neither mental, subjective things, nor are they antecedent entities like crude or raw material. They are the latter shaped for a purpose. It is impossible to overstate their objectivity from the standpoint of their existence and their efficacy within the operations in question; nor their objectivity in the sense of their dependence upon prior natural existences whose traits have to be taken account of, or reckoned with, by the operations of art. In the case of the art of inference, the art securely of going from the given to the absent, the dependence of mind upon inference, the fact that wherever inference occurs we have a conscious agent--one who recognizes, plans, invents, seeks out, deliberates, anticipates, and who, reacting to anticipations, fears, hates, desires, etc.--explains the theories which, because of misconception of the nature of mind and consciousness, have labeled logical distinctions psychical and subjective. In short, the theory shows why logical features have been made into ontological entities and into mental states.
To elaborate this thesis would be to repeat what has been said in all the essays of this volume. I wish only to call attention to certain considerations which may focus other discussions upon this hypothesis. 1. The existence of inference is a fact, a fact as certain and unquestioned as the existence of eyes or ears or the growth of plants, or the circulation of the blood. One observes it taking place everywhere where human beings exist. A student of the history of man finds that history is composed of beliefs, institutions, and customs which are inexplicable without acts of inference. This fact of inference is as much a datum--a hard fact--for logical theory as any sensory quality whatsoever. It is something men do as they walk, chew, or jump. There is nothing a priori or ideological about it. It is just a brute empirically observable event. 2. Its importance is almost as conspicuous as its existence. Every act of human life, not springing from instinct or mechanical habit, contains it; most habits are dependent upon some amount of it for their formation, as they are dependent upon it for their readaptation to novel circumstances. From the humblest act of daily life to the most intricate calculations of science and the determination and execution of social, legal, and political policies, things are used as signs, indications, or evidence from which one proceeds to something else not yet directly given.
3. The act of inferring takes place naturally, i.e., without intention. It is at first something we do, not something which we _mean_ to do. We do it as we breathe or walk or gesture. Only after it is done do we notice it and reflect upon it--and the great mass of men no more reflect upon it after its occurrence than they reflect upon the process of walking and try to discover its conditions and mechanism. That an individual, an animal organism, a man or a woman performs the acts is to say something capable of direct proof through appeal to observation; to say that something called mind, or consciousness does it is itself to employ inference and dubious inference. The fact of inference is much surer, in other words, than that of a particular inference, such as that to something called reason or consciousness, in connection with it; save as mind is but another word for the fact of inference, in which case of course it cannot be re-referred to as its cause, source, or author. Moreover, by all principles of science, inference cannot be referred to mind or consciousness as its condition, unless there is _independent_ proof of the existence of that mind to which it is referred.
Prima facie we are conscious or aware _of_ inference precisely as we are of anything else, not by introspection of something within the very consciousness which is supposed to be its source, but by observation of something taking place in the world--as we are conscious of walking _after_ we have walked. After it has been done naturally--or "unconsciously"--it may be done "consciously," that is, with intent or on purpose. But this means that it is done _with_ consciousness (whatever consciousness may be discovered to mean), not that it is done _by_ consciousness. Now if other natural events characteristic only (so far as can be ascertained) of highly organized beings are marked by unique or by distinctive traits, there is good ground for the assumption that inference will be so marked. As we do not find the circulation of blood or the stimulation of nerves in a stone, and as we expect as a matter of course to find peculiar conditions, qualities, and consequences in the being where such operations occur, so we do not find the act of inference in a stone, and we expect peculiar conditions, qualities, and consequences in whatever beings perform the act. Unless, in other words, all the ordinary canons of inquiry are suspended, inference is not an isolated nor a merely formal event. As against the latter, it has its own distinctive structure and properties; as against the former, it has specific generating conditions and specific results.
4. Possibly all this seems too obvious for mention. But there is often a virtual conspiracy in philosophy, not to mention obvious things nor to dwell upon them: otherwise remote speculations might be brought to a sudden halt. The point of these commonplaces resides in the push they may give anyone to engage in a search for _distinctive features in the act of inference_. The search may perhaps be best initiated by noting the seeming inconsistency between what has been said about inference as an art and inference as a natural, unpremeditated occurrence. The obvious function of spontaneous inference is to bring before an agent absent considerations to which he may respond as he otherwise responds to the stimulating force of the given situation. To infer rain is to enable one to behave _now_ as given conditions would not otherwise enable him to conduct himself. This instigation to behave toward the remote in space or time is the primary trait of the inferential act; descriptively speaking, the act consists in taking up an attitude of response to an absent thing as if it were present. But just because the thing is absent, the attitude taken may be either irrelevant and positively harmful or extremely pertinent and advantageous.
We may infer rain when rain is not going to happen, and acting upon the inference be worse off than if there had been no inference. Or we may make preparations, which we would not otherwise have made; the rain may come, and the inference save our lives--as the ark saved Noah. Inference brings, in short, truth and falsity into the world, just as definitely as the circulation of the blood brings its distinctive consequences, both advantages and liabilities into the world, or as the existence of banking brings with it consequences of business extension and of bankruptcy not previously existent. If the reader objects to the introduction of the terms "truth" and "falsity", I am perfectly willing to leave the choice of words to him, provided the fact is recognized that through inference men are capable of a kind of success and exposed to a kind of failure not otherwise possible: dependent upon the fact that inference takes absent things as being in a certain real continuum with present things, so that our attitude toward the latter is bound up with our reaction to the former as parts of the same situation. And in any event, I wish to protest against a possible objection to the introduction of the terms "false" and "true."
It may be said that inference is not responsible for the occurrence of errors and truths, because these accompany simple apprehensions where there is no inference: as when I see a snake which isn't there--or any other case which may appear to the objector to afford an illustration of his point. The objection illustrates my point. To affirm a snake is to affirm potentialities going beyond what is actually given; it says that what is given is _going_ to do something--the doing characteristic of a snake, so that we are to react to the given as to a snake. Or if we take the case of a face in the cloud recognized as a phantasy; then (to say nothing of "in the cloud" which involves reference beyond the given) "phantasy," "dream," equally means a reference to objects and considerations _not_ given as the actual datum is given. We have not got very far with our question of distinctive, unique traits called into existence by inference, but we have got far enough to have light upon what is called the "transcendence" of knowledge. All inference is a _going beyond_ the assuredly present to an absent. Hence it is a more or less precarious journey. It is transcending limits of security of immediate response. The stone which reacts only to stimuli of the present, not of the future, cannot make the mistakes which a being reacting to a future taken to be connected with the present is sure to make. But it is important to note just what this transcendence consists in. It has nothing to do with transcending mental states to arrive at an external object. _It is behaving to the given situation as involving something not given._ It is Robinson Crusoe going from a seen foot to an unseen man, not from a mental state to something unmental.
5. The mistakes and failures resulting from inference constitute the ground for transition from natural spontaneous performance to a technique or deliberate art of inference. There is something humorous about the discussion of the problem of error as if it were a rare or exceptional thing--an anomaly--when the barest glance at human history shows that mistakes have been the rule, and that truth lies at the bottom of a well. As to inferences bound up with barely keeping alive, man has had to effect a considerable balance of good guesses over bad. Aside from this somewhat narrow field, the original appearance of inference upon the scene probably added to the interest of life rather than to its efficiency. If the classic definition of man as a rational animal means simply an inferring or guessing animal, it applies to the natural man, for it allows for the guesses being mostly wrong. If it is used with its customary eulogistic connotations, it applies only to man chastened to the use of a hardly won and toilsome art. If it alleges that man has any natural preference for a reasonable inference or that the rationality of an inference is a measure of its hold upon him, it is grotesquely wrong. To propagate this error is to encourage man in his most baleful illusion, and to postpone the day of an effective and widespread adoption of a perfected art of knowing.
Summarily put, the waste and loss consequent upon the natural happening of inference led man, slowly and grudgingly, to the adoption of safeguards in its performance. In some part, the scope of which is easily exaggerated, man has come to attribute many of the ills from which he suffers to his own premature, inept, and unguarded performing of inference, instead of to fate, bad luck, and accident. In some things, and to some extent in all things, he has invented and perfected an art of inquiry: a system of checks and tests to be used before the conclusion of inference is categorically affirmed. Its nature has been considered in many other places in these pages, but it may prove instructive to restate it in this context.
When assured, such a simple datum is for the sake of guarding the act of inference. Color may mean a lot of things; any red may mean a lot of things; such things are ambiguous; they afford unreliable evidence or signs. To get the color down to the last touch of possible discrimination is to limit its range of testimony; ideally, it is to secure a voice which says but one thing and says that unmistakably. Its simplicity is not identical with isolation, but with _specified_ relationship. Thus the hard "facts," the brute data, the simple qualities or ideas, the sense elements of traditional and of contemporary logic, get placed and identified within the art of controlling inference. The allied terms "self-evident," "sensory truths," "simple apprehensions" have their meanings unambiguously determined in this same context; while apart from it they are the source of all kinds of error. They are no longer notions to conjure with. They express the last results attainable by present physical methods of discriminative analysis employed in the search for dependable data for inference. Improve the physical means of experimentation, improve the microscope or the registering apparatus or the chemical reagent, and they may be replaced tomorrow by new, simple apprehensions of simple and ultimate data.
While more careful inspection of the given object--to see if it be really a footprint, how fresh, etc.--may do much for safe-guarding inference; and while forays into whatever else is known may help, there is still need for something else. We need some method of freely examining and handling the object in its status as an inferred object. This means some way of detaching it, as it were, from the particular act of inference in which it presents itself. Without some such detachment, Crusoe can never get into a free and effective relation with the man indicated by the footprint. He can only, so to speak, go on repeating, with continuously increasing fright, "There's a man about, there's a man about." The "man" needs to be treated, not as man, but as something having a merely inferred and hence potential status; as a meaning or thought, or "idea." There is a great difference between meaning and _a_ meaning. Meaning is simply a function of the situation: this thing means that thing: meaning is this relationship. A meaning is something quite different; it is not a function, but a specific entity, a peculiar thing, namely the man _as_ suggested.
Words are the great instrument of translating a relation of inference existing between two things into a new kind of thing which can be operated with on its own account; the term of discourse or reflection is the solution of the requirement for greater flexibility and liberation. Let me repeat: Crusoe's inquiry can play freely around and about the man inferred from the footprint only as he can, so to say, get away from the immediate suggestive force of the footprint. As it originally stands, the man suggested is on the same coercive level as the suggestive footprint. They are related, tied together. But a gesture, a sound, may be used as a _substitute_ for the thing inferred. It exists independently of the footprint and may therefore be thought about and ideally experimented with irrespective of the footprint. It at once preserves the meaning-force of the situation and detaches it from the immediacy of the situation. It is a meaning, an idea.
Here we have, I submit, the explanation of notions, forms, essences, terms, subsistences, ideas, meanings, etc. They are surrogates of the objects of inference of such a character that they may be elaborated and manipulated exactly as primary things may be, so far as inference is concerned. They can be brought into relation with one another, quite irrespective of the things which originally suggested them. Without such free play reflective inquiry is mockery, and control of inference an impossibility. When a speck of light suggests to the astronomer a comet, he would have nothing to do but either to accept the inferred object as a real one, or to reject it as a mere fancy unless he could treat "comet" for the time being not as a thing at all, but as a meaning, a conception; a meaning having, moreover, by connection with other meanings, implications--meanings consequent from it. Unless a meaning is an inferred object, detached and fixed as a term capable of independent development, what sort of a ghostly Being is it? Except on the basis stated, what is the transition from the function of meaning to _a_ meaning as an entity in reasoning? And, once more, unless there is such a transition, is reasoning possible?
Cats have claws and teeth and fur. They do not have implications. No physical thing has implications. The _term_ "cat" has implications. How can this difference be explained? On the ground that we cannot use the "cat" object inferred from given indications in such a way as will test the inference and make it fruitful, helpful, unless we can detach it from its existential dependence upon the particular things which suggest it. We need to know what a cat would be _if_ it were there; what other things would also be indicated if the cat is really indicated. We therefore create a _new_ object: we take something to stand for the cat-in-its-status-as-inferred in contrast with the cat as a live thing. A sound or a visible mark is the ordinary mechanism for producing such a new object. Whatever the physical means employed, we now have a new object; a term, a meaning, a notion, an essence, a form or species, according to the terminology which may be in vogue. It is as much a specific existence as any sound or mark is. But it is a mark which notes, concentrates, and records an outcome of an inference which is not yet accepted and affirmed.
Ideas discussed in this chapter
Text from Project Gutenberg, public domain in the United States.