Original Conduit Metaphor Article

The text of my groundbreaking article on a coherent, underlying "folk model" for human communication found in the English language follows.  Comments are turned on here, so feel free to make them or ask questions.  You can also read a summary of this work in Wikipedia.  MR

The conduit metaphor: A case of frame conflict in our language about language.

by Michael J. Reddy

I should like to respond to Professor Schön’s chapter by replaying his theme several octaves lower. In my opinion, he has struck exactly the right set of notes. “Problem setting” should indeed be considered the crucial process, as opposed to “problem solving.” And the “stories that people tell about troublesome situations” do set up or “mediate” the problem. And “frame conflict” between various stories should be studied in detail, precisely because it is quite often “immune to resolution by appeal to the facts.”

It is hard to think of a better overture to genuine advance in the social and behavioral sciences than this. At the same time, it seems to me that Schön has managed to sound these excellent notes only in their overtones, so that the fundamental frequency is barely to be heard—even though, to my ears at least, Schön’s kind of thinking is real and long awaited music.

Quite simply, what I believe is missing is the application of Schön’s wisdom—this paradigm-consciousness—to human communication itself. It may seem predictable that I, a linguist, would take such a position. But, if I do, it is hardly disciplinary narrow-mindedness that motivates me. In 1954, Norbert Wiener, one of the originators of information theory, and the “father of cybernetics,” stated quite flatly: “Society can only be understood through a study of the messages and communications facilities which belong to it” (Wiener, 1954, p. 16).

I have never thought of this statement as referring to things like the size and adequacy of the telephone system. Wiener was talking primarily about the basic processes of human communication—how they work, what sort of wrinkles there are in them, when and why they are likely to succeed or fail. The problems of society, government, and culture depend ultimately on something like the daily box score of such successes or failures to communicate. If there are too many failures, or systematic types of failure, troubles will multiply. A society of near-perfect communicators, though it would no doubt still face conflicts of interest, might well be able to avoid many of the destructive, divisive effects of these inevitable conflicts.

The conduit metaphor

What do speakers of English say when communication fails or goes astray? Let us consider (1) through (3), some very typical examples,

   (1) Try to get your thoughts across better
   (2) None of Mary’s feelings came through to me with any clarity
   (3) You still haven’t given me any idea of what you mean

and do as Schön has suggested—take them as problem-setting stories, as descriptions of “what is wrong and what needs fixing.”

Are there metaphors in the examples? Do these metaphors set the directions for possible problem-solving techniques? Although (1) through (3) contain no fresh metaphors, there is in each case a dead metaphor. After all, we do not literally “get thoughts across” when we talk, do we? This sounds like mental telepathy or clairvoyance, and suggests that communication transfers thought processes somehow bodily. Actually, no one receives anyone else’s thoughts directly in their minds when they are using language.

Mary’s feelings, in example (2), can be perceived directly only by Mary; they do not really “come through to us” when she talks. Nor can anyone literally “give you an idea”—since these are locked within the skull and life process of each of us. Surely, then, none of these three expressions is to be taken completely at face value. Language seems rather to help one person to construct out of his own stock of mental stuff something like a replica, or copy, of someone else’s thoughts—a replica which can be more or less accurate, depending on many factors. If we could indeed send thoughts to one another, we would have little need for a communications system.

If there are dead metaphors in (1) through (3), then they all seem to involve the figurative assertion that language transfers human thoughts and feelings. Notice that this assertion, even in its present, very general form, leads already to a distinct viewpoint on communications problems. A person who speaks poorly does not know how to use language to send people his thoughts; and, conversely, a good speaker knows how to transfer his thoughts perfectly via language. If we were to follow this viewpoint, the next question would be: What must the poor speaker do with his thoughts if he is to transfer them more accurately by means of language?

The surprising thing is that, whether we like it or not, the English language does follow this viewpoint. It provides, in the form of a wealth of metaphorical expressions, answers to this and other questions, all of which answers are perfectly coherent with the assumption that human communication achieves the physical transfer of thoughts and feelings. If there were only a few such expressions involved, or if they were random, incoherent figures of speech arising from different paradigms—or if they were abstract, not particularly graphic images—then one might just succeed in dismissing them as harmless analogies. But in fact, none of these mitigating circumstances comes into play.

Typical solutions to the unskilled speaker’s communications problems are illustrated by (4) through (8).

   (4) Whenever you have a good idea practice capturing it in words
   (5) You have to put each concept into words very carefully
   (6) Try to pack more thoughts into fewer words
   (7) Insert those ideas elsewhere in the paragraph
   (8) Don’t force your meanings into the wrong words

Naturally, if language transfers thought to others, then the logical container, or conveyor, for this thought is words, or word-groupings like phrases, sentences, paragraphs, and so on.

One area of possible difficulty is then the insertion process. The speaker might be generally unpracticed or careless about this, and so be admonished with (4) or (5). As (6) shows, he could fail to put enough meaning in. Or, according to (7), he could put the right meanings in, but put them in the wrong place. Example (8), which stretches common sense most seriously, indicates that he might put meanings into the words which somehow do not fit in them, thus presumably deforming these meanings. It might also be, of course, that the speaker puts too much meaning into words. And there are expressions for this as well.

   (9) Never load a sentence with more thoughts than it can hold

In general, this class of examples implies that, in speaking or writing, humans place their internal thoughts and feelings within the external signals of the language. A more complete listing may be found in the Appendix.

The logic of the framework we are considering—a logic which will henceforth be called the conduit metaphor—would now lead us to the bizarre assertion that words have “insides” and “outsides.” After all, if thoughts can be “inserted,” there must be a space “inside” wherein the meaning can reside. But surely the English language, whatever metaphysical meanderings it may have been guilty of thus far, cannot have involved us in this kind of patent nonsense. Well, a moment’s reflection should nudge anyone into remembering that “content” is a term used almost synonymously with “ideas” and “meaning” And that recollection is quite meaning-full (sic) in the present context. Numerous expressions make it clear that English does view words as containing or failing to contain thoughts, depending on the success or failure of the speaker’s “insertion” process.

   (10) That thought is in practically every other word
   (11) The sentence was filled with emotion
   (12) The lines may rhyme, but they are empty of both meaning and
feeling
   (13) Your words are hollow—you don’t mean them

Or, in general, there is another class of examples that imply that words contain or convey thoughts and feelings when communication is successful. We assert, without batting an eyelash, that “the meaning is right there in the words.” Further instances are to be found in the Appendix.

It may be that the fault in a communication failure does not lie with the speaker. Perhaps, somehow, the listener has erred. In the framework of the conduit metaphor, the listener’s task must be one of extraction. He must find the meaning “in the words” and take it out of them, so that it gets “into his head.” Many expressions show that English does view the matter in this way.

   (14) Can you actually extract coherent ideas from that prose?
   (15) Let me know if you find any good ideas in the essay
   (16) I don’t get any feelings of anger out of his words

Curiously, my initial work on these expressions suggests that it is easier, when speaking and thinking in terms of the conduit metaphor, to blame the speaker for failures. After all, receiving and unwrapping a package is so passive and so simple—what can go wrong? A package can be difficult or impossible to open. But, if it is undamaged, and successfully opened, who can fail to find the right things in it? Thus, there are graphic and powerful expressions which blame particularly writers for making the package hard to open, as in (17) through (19).

   (17) That remark is completely impenetrable
   (18) Whatever Emily meant, it’s likely to be locked up in that cryptic
little verse forever
   (19) He writes sentences in such a way as to seal up the meaning in them

But, apart from readers and listeners “not paying attention to what’s there in the words,” the conduit metaphor offers little explanation for failing to “find” enough thoughts or the right thoughts in “what someone says.” Should someone discover too many thoughts, however, we have a wonderfully absurd expression faulting him for this.

   (20) You’re reading things into the poem

The power of the framework to enforce consistency of rationale even when the results are inane should be apparent here. We must see the reader as having surreptitiously made use of his power to insert thoughts into words when he should have restricted himself purely to extraction. He sneaked those thoughts into the words himself and then turned around and pretended that he found them there. Perhaps because the problem of too much meaning occurs more often in reading, we have never developed the corresponding expression for speaking—“hearing things into the poem.” Instead, we use “reading things into” for both modalities. Once again, further examples appear in the Appendix.

Perhaps we should pause at this point and set up some apparatus for generalizing what we have seen so far. It is not the numbered sentences above that are important, but rather the expressions in italics. These expressions could appear in many different utterances and take many different forms, and we have as yet no way of isolating what is crucial to them. Notice, for instance, that in every example there has been one word, such as “ideas,” or “thoughts,” or “meanings,” or “feeling,” which denotes internal conceptual or emotional material. Apart from what seem to be minor stylistic co-occurence restrictions, these and other terms like them can be substituted freely for one another. Thus, it is irrelevant to an example which one of these is present, and it would be helpful to have some abbreviation for the entire group.

Let us picture each person as having a “repertoire” of mental and emotional material. This will allow us to say that any term denoting a repertoire member, abbreviated “RM,” will fit, say, as object in (1) and produce an example utterance. Underlying (1), (2), and (3), then, are what we shall call “core expressions,” which can be written as follows.

   (21) getRMacross [underlying (1)]
   (22) RMcomes through (to someone) [underlying (2)]
   (23) give (someone) RM [underlying (3)]

The parentheses in (22) and (23) indicate optional complements. Examples (4) through (20), in addition to a term from the RM group, all contain another term, such as “word,” “phrase,” “sentence,” or “poem.” These words, in their basic senses at least, designate the external physical patterns of marks or sounds that do pass between speakers. Such energies, unlike the thoughts themselves, are received bodily, and are what information theorists would have called “signals.” If we adopt this generic name for the second group, and abbreviate it as “S,” then the core expressions for (4) through (6) are,

   (24) captureRMinS [underlying (4)]
   (25) putRMintoS [underlying (5)]
   (26) packRMintoS [underlying (6)]

In the Appendix, the core expression is always given first, and is then followed by one or two examples. Obviously, each core expression can be responsible for a very large number of different sentences.

The conduit metaphor, and the core expressions which embody it, deserve a great deal more investigation and analysis. My listing of the core expressions is most likely far from complete, and the logical reverberations of this paradigm affect both the syntax and the semantics of many words which are not themselves part of the core expressions. Later on, we shall focus on one such reverberation, which affects the entire S group. Apart from this, however, we shall have to be content to close the present discussion with a brief characterization of some further types of core expression.

Our examples thus far have been drawn from the four categories which constitute the “major framework” of the conduit metaphor. The core expressions in these categories imply, respectively, that: (1) language functions like a conduit, transferring thoughts bodily from one person to another; (2) in writing and speaking, people insert their thoughts or feelings in the words; (3) words accomplish the transfer by containing the thoughts or feelings and conveying them to others; and (4) in listening or reading, people extract the thoughts and feelings once again from the words.

Beyond these four classes of expressions, there are a good many examples which have different, though clearly related, implications. The fact that it is quite foreign to common sense to think of words as having “insides” makes it quite easy for us to abstract from the strict, “major” version of the metaphor, in which thoughts and emotions are always contained in something. That is, the major framework sees ideas as existing either within human heads or, at least, within words uttered by humans.

The “minor” framework overlooks words as containers and allows ideas and feelings to flow, unfettered and completely disembodied, into a kind of ambient space between human heads. In this case, the conduit of language becomes, not sealed pipelines from person to person, but rather individual pipes which allow mental content to escape into, or enter from, this ambient space. Again, it seems that this extension of the metaphor is aided by the fact that, somewhere, we are peripherally aware that words do not really have insides.

In any case, whatever the cause of the extension, there are three categories of expressions in the minor framework. The categories imply, respectively, that: (1) thoughts and feelings are ejected by speaking or writing into an external “idea space”; (2) thoughts and feelings are reified in this external space, so that they exist independent of any need for living human beings to think or feel them; (3) these reified thoughts and feelings may, or may not, find their way back into the heads of living humans. Some outstanding examples of minor framework expressions are, for the first category,
putRMdown on paper

   (27) Put those thoughts down on paper before you lose them!
pourRMout
   (28) Mary poured out all of the sorrow she had been holding in for so long.
getRMout
   (29) You should get those ideas out where they can do some good.

And for the second category,
RMfloat around

   (30) That concept has been floating around for decades.
RMfind way
   (31) Somehow, these hostile feelings found their way to the ghettos of Rome.
findRM EX LOC
   (32) You’ll find better ideas than that in the library.
   (33) John found those ideas in the jungles of the Amazon, not in some classroom.

(EX LOC here stands for any locative expression designating a place other than within human beings, that is, an external locative.)
And for the third category,
absorbRM

   (34) You have to “absorb” Aristotle’s ideas a little at a time.
RMgo over someone’s head
   (35) Her delicate emotions went right over his head.
getRMinto someone’s head
   (36) How many different concepts can you get into your head in one evening?

For further examples, see the Appendix.

The toolmakers paradigm

In order to investigate the effect of the conduit metaphor on the thought processes of speakers of English, we need some alternate way of conceiving of human communication. We require another story to tell, another model, so that the deeper implications of the conduit metaphor can be drawn out by means of contrast. Simply speaking, in order to engage in frame restructuring about human communication, we need first an opposing frame.

To begin this other story, I should like to suggest that, in talking to one another, we are like people isolated in slightly different environments. Imagine, if you will, for sake of the story, a huge compound, shaped like a wagon wheel.

Each pie-shaped sector of the wheel is an environment, with two spokes and part of the circumference forming the walls. The environments all have much in common with one another—water, trees, small plants, rocks, and the like—yet no two are exactly alike. They contain different kinds of trees, plants, terrain, and so on. Dwelling in each sector is one person who must survive in his own special environment. At the hub of the wheel there is some machinery which can deliver small sheets of paper from one environment to another. Let us suppose that the people in these environments have learned how to use this machinery to exchange crude sets of instructions with one another—instructions for making things helpful in surviving, such as tools, perhaps, or shelters, or foods, and the like.

But there is, in this story, absolutely no way for the people to visit each other’s environments, or even to exchange samples of the things they construct. This is crucial. The people can only exchange these crude sets of instructions—odd looking blueprints scratched on special sheets of paper that appear from a slot in the hub and can be deposited in another slot—and nothing more. Indeed, since there is no way to shout across the walls of the sectors, the people only know of one another’s existence indirectly, by a cumulative series of inferences. This part of the story, the “no visiting and no exchange of indigenous materials rule,” we shall call the postulate of “radical subjectivity.”

In the analogy, the contents of each environment, the “indigenous materials,” represent a person’s repertoire. They stand for the internal thoughts, feelings, and perceptions which cannot themselves be sent to anyone by any means that we know of. These are the unique materials with which each person must work if he is to survive. The blueprints represent the signals of human communication, the marks and sounds that we can actually send to one another. We shall have to ignore the question of how the system of instructions became established, even though this is an interesting part of the story. We shall simply assume that it has reached some sort of steady state, and shall watch how it functions.

Suppose that Alex has discovered an implement that is very useful to him. Say he has learned to build a rake and finds he can use it to clear dead leaves and other debris without damaging the living plants. One day Alex goes to the hub and draws as best he can three identical sets of instructions for fashioning this rake and drops these sets in the slots for Bob, Curt, and Don. As a result, three people struggling along in slightly different environments now receive these curious sheets of paper, and each one goes to work to try to construct what he can from them. Alex’s environment has a lot of wood in it, which is probably why he has leaves to rake in the first place.

Bob’s sector, on the other hand, runs more to rock, and Bob uses a lot of rock in his constructions. He finds a piece of wood for the handle, but begins to make the head of the rake out of stone. Alex’s original rake head was wood. But since it never occurred to him that anything but wood would be available or appropriate, he did not try to specify wood for the head in his instructions.

When Bob is about halfway finished with the stone rake head, he connects it experimentally to the handle and realizes with a jolt that this thing, whatever it is, is certainly going to be heavy and unwieldy. He ponders its possible uses for a time, and then decides that it must be a tool for digging up small rocks when you clear a field for planting. He marvels at how large and strong Alex must be, and also at what small rocks Alex has to deal with. Bob then decides that two large prongs will make the rake both lighter and better suited to unearthing large rocks.

Quite happy with both his double-bladed rock-pick and his new ideas about what this fellow Alex must be like, Bob makes three identical sets of instructions himself, for his rock-pick, and inserts them in the slots for Alex, Curt, and Don. Alex, of course, now assembles a rock-pick following Bob’s instructions, except that he makes it entirely of wood and has to change the design a little if a wooden, two-pronged head is to be strong enough. Still, in his largely rockless environment, he cannot see much use for the thing, and worries that Bob has misunderstood his rake. So he draws a second set of more detailed instructions for the rake head, and sends them out to everyone.

Meanwhile, over in another sector, Curt, who is particularly interested in clearing out a certain swamp, has created, on the basis of these multiple sets of instructions—the hoe. After all, when you are dealing with swamp grass and muck, you need something that will slice cleanly through the roots. And Don, from the same sets of instructions, has come up with a gaff. He has a small lake and fishes quite a bit. Although it would be interesting to get to know Curt and Don, the primary heroes of this story are Alex and Bob. We return now to them for the climax of the great rake conversation, in which, to everyone’s surprise, some real communication takes place.

Alex and Bob, who have had profitable interchanges in the past, and thus do not mind working quite hard at their communications, have been caught up in this rake problem for some time now. Their instructions simply will not agree. Bob has even had to abandon his original hypothesis that Alex is a huge man who has only small rocks to deal with. It just does not fit the instructions he is getting.

Alex, on his side, is getting so frustrated that he is ready to quit. He sits down near the hub and, in a kind of absent-minded display of anger, grinds two pebbles together. Suddenly he stops. He holds these rocks up in front of his eyes and seems to be thinking furiously. Then he runs to the hub and starts scribbling new instructions as fast as he can, this time using clever iconic symbols for rock and wood, which he hopes Bob will understand. Soon Alex and Bob are both ecstatic. All sorts of previous sets of instructions, not just about rakes, but about other things as well, now make perfect sense. They have raised themselves to a new plateau of inference about each other and each other’s environments.

For purposes of comparison, let us now view this same situation once again, as the conduit metaphor would see it. In terms of the radical subjectivist paradigm for human communication, what the conduit metaphor does is permit the exchange of materials from the environments, including the actual constructs themselves. In our story, we would have to imagine a marvelous technological duplicating machine located in the hub. Alex puts his rake in a special chamber, pushes a button, and instantly precise replicas of the rake appear in similar chambers for Bob, Curt, and Don to make use of. Bob, Curt, and Don do not have to construct anything or guess about anything.

Should Bob want to communicate with Curt and Don about Alex’s rake, there is no excuse for his sending anything except an exact replica of that rake to these people. There will still be differences in environments, but learning about these is now a trivial matter. Everything Bob has ever sent to Alex has been constructed largely of rock, and Alex is thus perfectly aware of his neighbor’s predicament. Even if the marvelous machine should falter now and again, so that artifacts arrive damaged, still, damaged objects look like damaged objects. A damaged rake does not become a hoe. One can simply send the damaged object back, and wait for the other person to send another replica.

It should be clear that the overwhelming tendency of the system, as viewed by the conduit metaphor, will always be: success without effort. At the same time, it should be similarly obvious that, in terms of the toolmakers paradigm, and the postulate of radical subjectivity, we come to just the opposite conclusion. Human communication will almost always go astray unless real energy is expended.

This comparison, then, brings to light a basic conflict between the conduit metaphor and the toolmakers paradigm. Both models offer an explanation of the phenomenon of communication. But they come to totally different conclusions about what, in that phenomenon, are more natural states of affairs, and what are less natural, or constrained, states. In terms of the conduit metaphor, what requires explanation is failure to communicate. Success appears to be automatic. But if we think in terms of the toolmakers paradigm, our expectation is precisely the opposite.

Partial miscommunication, or divergence of readings from a single text, are not aberrations. They are tendencies inherent in the system, which can only be counteracted by continuous effort and by large amounts of verbal interaction. In this view, things will naturally be scattered, unless we expend the energy to gather them. They are not, as the conduit metaphor would have it, naturally gathered, with a frightening population of wrong-headed fools working to scatter them.

As many scholars have pointed out (Kuhn, 1970a; Butterfield, 1965), such shifts in the notion of what a thing does “naturally,” that is, if left to its own devices, are the stuff of which scientific revolutions are made. If the earth holds still at some center point, then it is the motions of celestial bodies that must be theorized about and predicted. But if the sun is at that center point, then we must theorize about the motion of the earth.

In this regard, the present situation is a little curious. The toolmakers paradigm is very much in accord with the long-postulated connection between information, in the mathematical sense, and the entropy expression of the second law of thermodynamics (Cherry, 1966, pp. 214–17). The second law states that if left to their own devices, all forms of organization always decrease in time. Successful human communication involves an increase in organization, which cannot happen spontaneously or of its own accord.

Thus, the shift in viewpoint of the toolmakers paradigm merely seems to bring the model of human communication into line with a previously extant paradigm from the physical sciences. But even though, mathematically, information is expressed as negative entropy, debate and confusion have always surrounded this connection. And it may be that this confusion springs, in part at least, from the dominant position occupied by the conduit metaphor in our language. For the conduit metaphor is definitely in conflict with the second law.

But I do not want to argue too strongly either for or against either of these models in this paper. I do not want to attempt any “appeal to the facts” at this point. For the real question here is to what extent language can influence thought processes. To me, from my vantage point now, it seems that the toolmakers paradigm and radical subjectivism simply form a coherent, common-sense view of what happens when we talk—a common-sense view which finds support in everything from this second law of thermodynamics to recent work in artificial intelligence or cognitive psychology. But if my major claim is true—that the conduit metaphor is a real and powerful semantic structure in English, which can influence our thinking—then it follows that “common sense” about language may be confused.

I confess that it took nearly five years for me to come around to radical subjectivism as “common sense.” What stood in the way was never a counter-argument, but rather the simple inability to think clearly about the matter. My mind would seem to go to sleep at crucial moments, and it was only the mounting weight of more and more evidence that finally forced it to stay awake. Thus, there is the likelihood that arguments about these models either will not be needed, or alternatively, if they are needed, will fall on deaf ears until the biasing effect of the conduit metaphor has been dealt with. Most important, then, is some survey of the evidence that the conduit metaphor can and does influence our thinking.

Semantic pathology

Let us assume now, for the sake of argument, that it is agreed that communication functions as the toolmakers paradigm suggests, and not as the conduit metaphor would have it. And let us assume further that the conflicting implications of the two frames are theoretically interesting or even important. You may well grant me these things and still hold that the conduit metaphor expressions in everyday language do not really influence, or confuse, our thought processes.

After all, all of us succeeded in shifting mental gears and thinking about language in terms of the toolmakers paradigm right here in the present discussion. The conduit metaphor did not prevent us from doing this. Where really is the problem? How can anything troublesome arise from a conceptual frame that we were able to discard so easily? This is the question to which we shall address ourselves now. Can the conduit metaphor really bias our thinking? And if so, how?

To begin with, it must be made clear that no speaker of English, not even your author, has discarded the conduit metaphor. Thinking in terms of the toolmakers paradigm briefly may, perhaps, have made us aware of the conduit metaphor. But none of us will discard it until we succeed in bringing about an entire series of linked changes in the English language.

The logic of the framework runs like threads in many directions through the syntactic and semantic fabric of our speech habits. Merely becoming cognizant of this in no way alters the situation. Nor does it appear that one can adopt a new framework and develop it while ignoring the cloth of the language. For everywhere one runs into the old threads, and each one pushes conversation and thought back a little way toward the established pattern. No matter how otherworldly this may seem, there is some exceedingly poignant evidence that it has occurred and continues to occur.

The precise claim being made here is important. It has to do, I think, with one of the ways in which people commonly misunderstand the Whorf hypothesis. I do not claim that we cannot think momentarily in terms of another model of the communication process. I argue, rather, that that thinking will remain brief, isolated, and fragmentary in the face of an entrenched system of opposing attitudes and assumptions.

I have not been able to gather hard statistics about the number of core expressions arising from the conduit metaphor. Indeed, inasmuch as the concept of a “core expression” is itself somewhat loose, and inasmuch as it is difficult in some cases to decide whether an expression should or should not be listed, I am not sure whether hard statistics can ever be assembled. Nevertheless, the present tally of conduit metaphor expressions is about 140.

If one looks about for alternative ways of speaking about communication—ways which are either metaphorically neutral, or metaphorically opposed to the conduit framework—the list of expressions numbers between 30 and 40. A conservative estimate would thus be that, of the entire metalingual apparatus of the English language, at least seventy percent is directly, visibly, and graphically based on the conduit metaphor.

Whatever influence the remaining thirty percent might have appears to be weakened beyond this direct proportionality by several factors. First, these expressions tend to be the multisyllabic, latinate abstractions (“communicate, “disseminate” “notify, “disclose” and so on) which are neither graphic nor metaphorically coherent. Thus, they do not present an alternative model of the communication process, which leaves the notion of  “putting ideas into words” as the sole available conception.

Second, most of them can be used with the adjunct “in words” (“in S,” more generally), thereby losing their neutrality and lending added support to the conduit metaphor. “Communicate your feelings using simpler words,” for example, succeeds in avoiding the conduit metaphor, whereas, “Communicate your feelings in simpler words,” does not. And finally, to the extent that etymologies are relevant, many of these expressions have roots which spring directly from the conduit framework (“express,” “disclose,” etc). See Part Two of the Appendix for this listing.

The simplest, and perhaps most convincing illustration of our dependence on the conduit metaphor core expressions is a test that can be performed by anyone. Familiarize yourself with the listings in the Appendix. Then begin to become aware of, and try to avoid, conduit metaphors. Every time you find yourself using one, see if you can replace it with a neutral expression, or some circumlocution. My experience in teaching classes which dealt with this subject has been that I am constantly called to account by my students for using the expressions I am lecturing about. If I speak very carefully, with constant attention, I can do fairly well at avoiding them. But the result is hardly idiomatic English. Instead of walking into a classroom and asking “Did you get anything out of that article?” I have to say, “Were you able to construct anything of interest on the basis of the assigned text?”

If one should look, I daresay even the present article is not free from conduit metaphor expressions. I ended the preceding section with a minor framework, category three example, (141) in the Appendix, when I wrote: “The arguments will fall on deaf ears.” Practically speaking, if you try to avoid all obvious conduit metaphor expressions in your usage, you are nearly struck dumb when communication becomes the topic. You can say to your wayward student, “Try to communicate more effectively, Reginald,” but it will not have nearly the impact of, “Reginald, you’ve got to learn how to put your thoughts into words.”

But even if you could avoid all such obvious conduit “metaphorisms,” this would still not free you from the framework. The threads, as I said, are nearly everywhere. To see that they go much deeper than just a list of expressions, I should like to resurrect a concept from pretransformational semantics. In his Principles of Semantics, Stephen Ullmann (1957, p. 122) makes use of the term semantic pathology. A semantic pathology arises “whenever two or more incompatible senses capable of figuring meaningfully in the same context develop around the same name.”

For some time, my favorite English illustration of this was the delicate and difficult problem of distinguishing sympathy from apology. That is, “I’m sorry,” can mean either “I empathize with your suffering,” or “I admit fault and apologize.” Sometimes people expect apologies from us when we only wish to sympathize, in which case saying, “I’m sorry,” is either the perfect hedge or the opening line of a fight. Other times, people think we are apologizing when they see no need for us to apologize and respond with, “That’s alright, it wasn’t your fault.”

As I studied the conduit metaphor, however, I came to rely on this example less and less. I kept coming across terms which were ambiguous between what we have here called “repertoire members” and what we have called “signals.” I would find a word which, in its basic sense, referred to some grouping of the marks or sounds which we do exchange with one another. But then I would use it in sentences and realize that it could refer just as easily and just as often to segments of human thought or emotions. Consider the word “poem,” for example. In (37) through (39),

   (37) The poem was almost illegible
   (38) The poem has five lines and forty words
   (39) The poem is unrhymed

This word clearly refers to a text, some signals involving either marks or sounds. For sake of clarity, let us call the word-sense operating here POEM1, (for an operational definition of “word-sense,” see Reddy, 1973). Now notice that, in (40) through (42),

   (40) Donne’s poem is very logical
   (41) That poem was so completely depressing
   (42) You know his poem is too obscene for children

the most probable referent of the work is not a text, but rather the concepts and emotions assembled in the reading of a text. I say “most probable” here because it is possible to imagine contexts in which the referent is actually once again a text. Suppose, for instance, (41) is uttered by a teacher of penmanship about a child’s hasty copy of some poem. Barring such unusual contexts, however, “poem” in these examples refers to conceptual and emotional material. The word-sense functioning here we shall call POEM1. Example (43) can be read with eitherPOEM1 or POEM2.

   (43) Martha’s poem is so sloppy!

It is easy to see that this ambiguity of the term “poem” is intimately related to the conduit metaphor. If the words in language contain the ideas, then POEM1 contains POEM2, and metonymy, a process of meaning extension second in importance only to metaphor, takes over. That is, when two entities are always found together in our experience, the name of one of them—usually the more concrete—will develop a new sense which refers to the other. Just as ROSE1 (= the blossom) developed ROSE2 (= the shade of pinkish red) by metonymy, so POEM1 gave rise to POEM2. For, in terms of the conduit metaphor, the two are seen as existing together, the second within the first, and all the conditions for metonymy are met. As long as we are happy with the conduit metaphor, then this ambiguity is in no way problematic, and is certainly not a semantic pathology.

But now consider what happens to the linguistic idealist who wants to think about communication in terms of the toolmakers paradigm and radical subjectivism without making any changes in the English language. In this new model , the words do not contain the ideas, and so POEM1 does not contain POEM2. Instead, it is of greatest importance to preserve a principled distinction between POEM1 and POEM2. There is in most cases only one POEM1, one text, to worry about.

But because of the differences in repertoires from one person to the next, and because of the difficult task of assembling these mental and emotional materials on the basis of the instructions in the text, it is obvious to our theorist that there will be as many POEM2’s in existence as there are readers or listeners. These internal POEM2’s will only come to resemble one another after the people expend some energy talking with one another and comparing notes. There is now not the slightest basis for a metonymical extension of POEM1 to POEM2.

If we had viewed language in terms of the toolmakers paradigm historically, these two profoundly different concepts would never have been accessed by the same word. Talking about an entire series of slightly, or even terribly, different entities as if there were only one would obviously have led to communicative disaster.

We see, then, that things have taken a troublesome turn for our linguistic idealist. This ambiguity of the word “poem” is for him a real and severe semantic pathology. Other speakers, who accept the conduit metaphor, can be perfectly blase about it. But he cannot. It befuddles the very distinction he is most concerned to make and bring others to make. More troublesome still is the fact that this pathology is global. It is not an isolated development in the language, involving only the word “poem.” I have discussed “poem” here as a paradigm case for the entire class of words in English which denote signals. Analogous examples are available for all of the S words discussed on page 169—“word,” “phrase,” “sentence,” “essay,” “novel,” and so on. Even the word “text” has the two senses, as evidenced by (44) and (45):

   (44) I am tired of illegible texts
   (45) The text is logically incoherent

In addition, all the proper names of texts, poems, plays, novels, speeches, and the like share this ambiguity. Notice,

   (46) The Old Man and the Sea is 112 pages long
   (47) The Old Man and the Sea is deeply symbolic

As I became aware of this systematic, widespread semantic pathology, I was, of course, far less impressed with the difficulties caused by, “I’m sorry.” For here was a case that involved more words than any pathology I had ever heard of. Furthermore, this case showed that semantic structures could be completely normal with respect to one view of reality, and at the same time, pathological with respect to another view. Or in other words, here was some strong evidence that language and views about reality have to develop hand in hand. Finally, I also noticed that this new, potential pathology affected what might be called the “morphosemantics” of the words involved. Suppose, for example, we pluralize the word “poem.” As shown in,

   (48) We have several poems to deal with today

This produces a form whose most natural referents are a number of POEM1’s, that is, a series of different texts.

It would be quite unnatural to utter (48) and mean that there were several internal POEM2’s, Michael’s POEM2, Mary’s POEM2, Alex’s POEM2, and so on, all constructed from the same POEM1, which were to be discussed on a given day. What this means is that, although POEM1, pluralizes with the change inmorphology, the other sense, POEM2, is lost in this change. In the case of proper names, pluralization is even more problematic. For most names of texts, there is no morphology defined for the plural. How should our budding radical subjectivist pluralize The Old Man and the Sea? Does he say, “Our internal The Old Man and the Sea-s”? Or should it be, “Our internal Old Man and the Sea”? And notice that it will not help him very much to use (49), or (50).

   (49) Our versions of the poem
   (50) Our versions of The Old Man and the Sea

For if, in (49), the word “poem” means POEM1, then this phrase applies to variants of the text—which is not what he wants to say. On the other hand, if “poem” means POEM2, then he is still in trouble. Now it sounds like there is one proper and correct POEM2, available to us all, which we may however, for reasons of taste, alter slightly. The radical subjectivism, the absolute nontransferability of any “correct” POEM2, is muddied completely by (49) and (50). This most important fact, that there is one POEM1 but necessarily many POEM2’s, cannot be expressed easily, consistently, or at all naturally.

This discussion, though it says by no means all that could be said, provides an initial illustration of what would happen to someone who really tried to discard the conduit metaphor and think seriously and coherently in terms of the toolmakers paradigm. He would face serious linguistic difficulties, to say the least, and would quite clearly have to create new language as he restructured his thought. But, of course, he would be likely to do this only if he shared our present awareness of the biasing power of the conduit metaphor. So far as I know, none of the thinkers who have tried to present alternate theories of language and the nature of meaning have had this awareness. Thus, the conduit metaphor has undercut them, without any knowledge on their part of what was happening.

Of course, the problems caused by this confusion in aesthetics and criticism are legion, and it is easy to document my claims by analysis of works in this area. However, a more convincing documentation—indeed, the most convincing documentation one could wish for—is to be found in the historical development of mathematical information theory. For here, if ever, with both a concept-free algebra of information, and working machines to use as models, the effect of the conduit metaphor should have been avoided. But, in fact, it was not. And the conceptual basis of the new mathematics, though not the mathematics itself, has been completely obscured by the semantic pathologies of the conduit metaphor.

The framework of mathematical information theory has much in common with our toolmakers paradigm. Information is defined as the ability to make nonrandom selections from some set of alternatives. Communication, which is the transfer of this ability from one place to another, is envisioned as occurring in the following manner. The set of alternatives and a code relating these alternatives to physical signals are established, and a copy of each is placed at both the sending and receiving ends of the system.

This act creates what is known as an “a priori shared context,” a prerequisite for achieving any communication whatsoever. At the transmitting end, a sequence of the alternatives, called the message, is chosen for communication to the other end. But this sequence of alternatives is not sent. Rather, the chosen alternatives are related systematically by the code to some form of energy patterns which can travel quickly and retain their shape while they do travel—that is, to the signals.

The whole point of the system is that the alternatives themselves are not mobile, and cannot be sent, whereas the energy patterns, the “signals” are mobile. If all goes well, the signals, when they arrive at the receiving end, are used to duplicate the original selection process and recreate the message. That is, using the relationships of the code and the copy of the original set of alternatives, the receiving end can make the same selections that were made earlier on the transmitting end when the message was generated. Quantification is possible in this framework only because one can set up measures of how much the received signals narrow down the possible choices of pre-existent alternatives.

In terms of our toolmakers paradigm, the predefined set of alternatives of information theory corresponds to what we have called the “repertoire.” The environments of the persons in the wagon-wheel compound all have much in common—otherwise their system of instructions would not work at all. The “signals” of the mathematical theory are exactly the same as our “signals”—the patterns that can travel, that can be exchanged. In the world of the compound, they are the sheets of paper sent back and forth.

Notice, now, that in information theory, as in our paradigm, the alternatives—the “messages”—are not contained in the signals. If the signals were to arrive at the receiving end, and the set of alternatives was damaged or missing, the proper selections could not be made. The signals have no ability to bring the alternatives with them; they carry no little replica of the message. The whole notion of information as “the power to make selections” rules out the idea that signals contain the message.

Now, this may be abundantly clear when spelled out in this fashion. And it seems to remain clear as long as information theory is restricted to simple, technical applications. But as most of you know, this theory was hailed as a potential breakthrough for biology and the social sciences. And numerous attempts were made to extend its range of application to include human language and behavior (see Cherry, 1966). Such attempts, of course, were not simple and technical.

They required a very clear understanding, not so much of the mathematics of the theory, but rather of the conceptual foundations of the theory. By and large, these attempts were all accounted to be failures. I think that the reason for these failures was the interaction of the conduit metaphor with the conceptual foundations of information theory. As soon as people ventured away from the original, well-defined area of the mathematics, and were forced to rely more on ordinary language, the essential insight of information theory was muddled beyond repair.

The destructive impact of ordinary language on any extensions of information theory begins with the very terms the originators (Shannon & Weaver, 1949) chose to name parts of the paradigm. They called the set of alternatives, which we have referred to here as the “repertoire,” the alphabet. It is true that in telegraphy the set of alternatives is in fact the alphabet; and telegraphy was their paradigm example. But they made it quite clear that the word “alphabet” was for them a technical coinage which was supposed to refer to any set of alternative states, behaviors, or what have you. But this piece of nomenclature is problematic when one turns to human communication.

For years I taught information theory in a nonmathematical way to future English teachers, using the term “alphabet.” Always this seemed to confuse them, though I never could fathom why, until one year, a student put up her hand and said, “But you can’t call the alternatives the signals.” Now it is strange, on the face of it, that Weaver, particularly, who was very concerned about applying the theory to human communication, would have let this go unnoticed. It confuses the all-important distinction between signals and repertoire members. Substituting the present term, “repertoire,” for “alphabet” made my teaching much easier.

But another mistake in terminology makes it seem probable that Shannon and Weaver were never quite clear themselves about the importance of this distinction to their own system. Consider the choice of the term “message” to represent the selection of alternatives from the repertoire. “Message,” as the following examples show, partakes of the same semantic pathology as “poem.”

   (51) I got your message (MESSAGE1), but had no time to read it
   (52) Okay, John, I get the message (MESSAGE2); let’s leave him alone

For information theory, this is extremely confusing, because MESSAGE1 means literally a set of signals, whereas MESSAGE2 means the repertoire members involved with the communication. For conduit-metaphor thinking, in which we send and receive the MESSAGE2, within the MESSAGE1, the ambiguity is trivial. But for a theory based totally on the notion that the “message” (MESSAGE2) is never sent anywhere, this choice of words leads to the collapse of the paradigm.

Shannon and Weaver were very careful to point out that the “received signals” were not necessarily the “transmitted signal” because of the possible intervention of distortion and noise. But they blithely wrote the word “message” on the right, or receiving side of their famous paradigm (Shannon & Weaver, 1949, p. 7). At the very least they should have written “reconstructed message” there. In their theory, something is rebuilt on that right side which, hopefully, resembles the original message on the left side. The ambiguity of the word “message” should have led them to regard this word as a disaster and never to consider it for use.

If they did not, I believe it is because their thought processes were responding to the biasing effect of the conduit metaphor. Weaver, it seems, could not hold the theory clearly in mind when he spoke of human communication, and used conduit metaphor expressions almost constantly. “How precisely,” he asked, “do the transmitted symbols convey the desired meaning?” [italics mine] (p. 4). Or he compared two “messages, one of which is heavily loaded with meaning and the other of which is pure nonsense” (p. 8). In truth, it seems that he still thought of the MESSAGE2, the repertoire members, as being sent across the channel, even though this destroys the notion of information as selective power.

Weaver hedges significantly when he describes the action of the transmitter. It “changes” he says, “the message into the signal” [italics Weaver’s] (p. 7). Really, this is a strange description. A code is a relationship between two distinct systems. It does not “change” anything into anything else. It merely preserves in the second system the pattern of organization present in the first system. Marks or sounds are not transmuted into electronic pulses. Nor are thoughts and emotions magically metamorphosed into words. Again, this is conduit-metaphor thinking. There is no justification whatsoever in information theory for talking about communication this way.

It is worth noting that Shannon, who actually originated the mathematics, may have had a more coherent understanding than Weaver. At some points in his own exposition, Shannon used exactly the right ordinary language terms. He wrote, “The receiver ordinarily performs the inverse operation of that done by the transmitter, reconstructing the message from the signal” (p. 34). But it still does not seem that he perceived the damage done to the paradigm by his own and Weaver’s conduit metaphorisms.

Quite the same thing can be said for other ways of speaking associated with information theory. They do violence to the theory, yet support and uphold the conduit metaphor admirably. Consider “encode” and “decode.” These mean to put the repertoire members “into” code, and then take them “out of” code, respectively. Or think about the term “information content.” The theory conceives of information as the power to reproduce an organization by means of nonrandom selections. Signals do something. They cannot contain anything.

If the conduit metaphor is capable of influencing thought processes, then why has an entire generation of information theorists talked in this confusing and detrimental way? One would have to suppose that Weaver and many researchers who have followed him were simply bent on professional destruction. It seems easier to believe that the English language has the power to lead them astray. A recent anthology collecting psychological and sociological efforts to create a communication theory for human interactions points out in the introduction that “investigators have yet to establish a completely acceptable definition of communication” (Sereno & Mortensen, 1970, p. 2). Then it goes on to say,

Those models based upon a mathematical conception describe communication as analogous to the operations of an information processing machine: an event occurs in which a source or sender transmits a signal or message through a channel to some destination or receiver. [italics from anthology] (p. 71)

Notice the statement, “transmits a signal or message.” Here, twenty-one years after Shannon and Weaver, the same confusion persists—can the “message” be sent, or not? And it persists in almost every article of the volume. Consider one more brief example. “The theory [of information] was concerned with the problem of defining the quantity of information contained in a message to be transmitted...” (p. 62). Note that here information is contained in a transmitted “message.” If the author means MESSAGE1, then he is thinking in terms of the conduit metaphor, and saying that information is contained in the signals. If he means MESSAGE2, then he is saying that repertoire members, which are transmitted inside of signals, have inside of them something called information, which can be measured. Either way, the insight of information theory has been overwhelmed.

Social implications

I should like to conclude with some remarks on the social implications of the situation we have outlined. If the English language has a less than accurate idea of its own workings, and if it has the power to bias thought processes in the direction of this model, what practical impact does this have? We have seen evidence that the conduit metaphor can confuse serious attempts at theory building—but does it matter at all to the man on the street, to mass culture, to federal policy-making?

I must limit myself here to suggesting two ways in which the conduit metaphor does matter to all speakers of English. To discuss the first way, I would like to return to the “stories” told in an earlier section and add a final sequel.

It came to pass, one year, that an evil magician, who was an expert at hypnosis, flew over the toolmakers’ compound. Looking down, he saw that, despite the formidable handicaps, Alex, Bob, Curt, and Don were doing quite well with their system of instruction sending. They were very aware that communicating was hard work. And their successes were extremely rewarding to them, because they retained a distinct sense of awe and wonder that they could make the system work at all. It was a daily miracle, which had improved their respective standards of living immensely.

The evil magician was very upset about this, and decided to do the worst thing he could think of to Alex, Bob, Curt, and Don. What he did was this. He hypnotized them in a special way, so that, after they received a set of instructions and struggled to build something on the basis of them, they would immediately forget about this. Instead, he planted in them the false memory that the object had been sent to them directly from the other person, via a marvelous mechanism in the hub. Of course, this was not true. They still had to build the objects themselves, out of their own materials—but the magician blinded them to this.

As it turned out, the evil magician’s shrewdness was profound. For even though, objectively, the communications system of the compound had not changed one bit, it nevertheless fell very quickly into disuse and decay. And as it crumbled, so did the spirit of harmony and communal progress that had always characterized the relations of Alex, Bob, Curt, and Don. For now, since they would always forget that they had assembled an object themselves and thus bore a large share of responsibility for its shape, it was easy to ridicule the sender for any defects.

They also began to spend less and less time working to assemble things, because, once the mental block descended, there was no feeling of reward for a job well done. As soon as they finished an assembly, the hypnosis would take effect, and suddenly—well, even though they were worn out, still, it was the other fellow who had done all the hard, creative work of putting it together. Any fool could take a finished product out of the chamber in the hub. So they came to resent, and therefore abandon, any assembly jobs that required real work.

But this was not the worst effect foreseen by the evil magician when he cast his peculiar spell. For, indeed, it was not long before each of them came to entertain, privately, the idea that all the others had gone insane. One would send instructions to the others for some device of which he was particularly proud, just as he had always done. Only now, of course, he believed that he sent, not instructions, but the thing itself. Then, when the others would send him instructions in return, to confirm their receipt of his, he would assemble the object, forget, think that they had returned him the thing itself, and then stare in horror at what he saw.

Here he had sent them a wonderful tool, and they returned to him grotesque parodies. Really, what could explain this? All they had to do was to successfully remove his object from the chamber in the hub. How could they change it so shockingly in performing an operation of such moronic simplicity? Were they imbeciles? Or was there perhaps some malice in their behavior? In the end, Alex, Bob, Curt, and Don all came privately to the conclusion that the others had either become hostile or else gone berserk. Either way, it did not matter much. None of them took the communications system seriously any more.

Among other things, this sequel attempts to sketch some of the social and psychological effects of believing that communication is a “success without effort” system, when, in fact, it is an “energy must be expended” system. I am sure that no one has failed to realize that, to the extent that the parable applies, the evil magician is the English language, and his hypnotic spell is the bias imparted to our thought processes by the conduit metaphor.

This model of communication objectifies meaning in a misleading and dehumanizing fashion. It influences us to talk and think about thoughts as if they had the same kind of external, intersubjective reality as lamps and tables. Then, when this presumption proves dramatically false in operation, there seems to be nothing to blame except our own stupidity or malice. It is as if we owned a very large, very complex computer—but had been given the wrong instruction manual for it. We believe the wrong things about it, and teach our children the wrong things about it, and simply cannot get full or even moderate usage out of the system.

Another point from the story worth emphasizing is that, to the extent that the conduit metaphor does see communication as requiring some slight expenditure of energy, it localizes this expenditure almost totally in the speaker or writer. The function of the reader or listener is trivialized. The radical-subjectivist paradigm, on the other hand, makes it clear that readers and listeners face a difficult and highly creative task of reconstruction and hypothesis testing. Doing this work well probably requires considerably more energy than the conduit metaphor would lead us to expect.

But we are still a long way from government policy in these effects. Let us turn, then, to the second example of the impact of the conduit metaphor, which will help to close this gap. The expression employed in (53), number 114 in the Appendix,

   (53) You’ll find better ideas than that in the library,

is derived from the conduit metaphor by a chain of metonymies. That is, we think of the ideas as existing in the words, which are clearly there on the pages. So the ideas are “there on the pages” by metonymy. Now the pages are in the books—and again, by metonymy, so are the ideas. But the books are in the libraries, with the final result that the ideas, too, are “in the libraries.” The effect of this, and the many other minor framework core expressions is to suggest that the libraries, with their books, tapes, films, and photographs, are the real repositories of our culture. And if this is true, then naturally we of the modern period are preserving our cultural heritage better than any other age, because we have more books, films, tapes, and so on, stored in more and bigger libraries.

Suppose now that we drop the conduit metaphor and think of this same situation in terms of the toolmakers paradigm. From this point of view, there are of course no ideas in the words, and therefore none in any books, nor on any tapes or records. There are no ideas whatsoever in any libraries. All that is stored in any of these places are odd little patterns of marks or bumps or magnetized particles capable of creating odd patterns of noise. Now, if a human being comes along who is capable of using these marks or sounds as instructions, then this human being may assemble within his head some patterns of thought or feeling or perception which resemble those of intelligent humans no longer living.

But this is a difficult task, for those no longer living saw a different world from ours, and used slightly different language instructions. Thus, if this human who enters the library has not been schooled in the art of language, so that he is deft and precise and thorough in applying instructions, and if he does not have a rather full and flexible repertoire of thoughts and feelings to draw from, then it is not likely that he will reconstruct in his head anything that deserves to be called “his cultural heritage.”

Quite obviously, the toolmakers paradigm makes it plain that there is no culture in books or libraries, that, indeed, there is no culture at all unless it is reconstructed carefully and painstakingly in the living brains of each new generation. All that is preserved in libraries is the mere opportunity to perform this reconstruction. But if the language skills and the habit of engaging in reconstruction are not similarly preserved, then there will be no culture, no matter how large and complete the libraries may become. We do not preserve ideas by building libraries and recording voices. The only way to preserve culture is to train people to rebuild it, to “regrow” it, as the word “culture” itself suggests, in the only place it can grow—within themselves.

The difference of viewpoint here between the conduit metaphor and the toolmakers paradigm is serious, if not profound. Humanists appear to be dying these days, and administrators and governments seem to feel few compunctions about letting this occur. We have the greatest, most sophisticated system for mass communication of any society that we know about, yet somehow mass communication becomes more and more synonymous with less communication. Why is this? One reason, at least, may be that we are following our instruction manual for use of the language system quite carefully—and it is the wrong manual.

We have the mistaken, conduit metaphor-influenced view that the more signals we can create, and the more signals we can preserve, the more ideas we “transfer” and “store.” We neglect the crucial human ability to reconstruct thought patterns on the basis of signals and this ability founders. After all, “extraction” is a trivial process, which does not require teaching past the most rudimentary level. We have therefore, in fact, less culture—or certainly no more culture—than other, less mechanically inclined, ages have had.

Humanists, those traditionally charged with reconstructing culture and teaching others to reconstruct it, are not necessary in the scheme of the conduit metaphor. All the ideas are “there in the library,” and anyone can go in and “get them.” In the toolmakers paradigm, on the other hand, humanists themselves are the repositories, and the only real repositories of ideas. In the simplest of terms, the conduit metaphor lets human ideas slip out of human brains, so that, once you have recording technologies, you do not need humans any more.

I am suggesting, then, that in the same way that “urban renewal” misled the policymakers discussed in Schön’s paper, the conduit metaphor is leading us down a technological and social blind alley. That blind alley is mass communications systems coupled with mass neglect of the internal, human systems responsible for nine-tenths of the work in communicating. We think we are “capturing ideas in words,” and funneling them out to the greatest public in the history of the world. But if there are no ideas “within” this endless flood of words, then all we are doing is replaying the myth of Babel—centering it, this time, around a broadcasting tower.