In linguistics, the conduit metaphor is a dominant class of figurative expressions used when discussing communication itself (metalanguage). It operates whenever people speak or write as if they "insert" their mental contents (feelings, meanings, thoughts, concepts, etc.) into "containers" (words, phrases, sentences, etc.) whose contents are then "extracted" by listeners and readers. Thus, language is viewed as a "conduit" conveying mental content between people.
The conduit metaphor was first defined and described by linguist Michael J. Reddy in 1979.[1] Reddy's proposal of this conceptual metaphor refocused debate within and outside the linguistic community on the importance of metaphorical language.
Fellow linguist George Lakoff stated that
"The contemporary theory that metaphor is primarily conceptual, conventional, and part of the ordinary system of thought and language can be traced to Michael Reddy’s now classic essay... With a single, thoroughly analyzed example, he allowed us to see, albeit in a restricted domain, that ordinary everyday English is largely metaphorical, dispelling once and for all the traditional view that metaphor is primarily in the realm of poetic or 'figurative' language. Reddy showed, for a single, very significant case, that the locus of metaphor is thought, not language, that metaphor is a major and indispensable part of our ordinary, conventional way of conceptualizing the world, and that our everyday behavior reflects our metaphorical understanding of experience. Though other theorists had noticed some of these characteristics of metaphor, Reddy was the first to demonstrate them by rigorous linguistic analysis, stating generalizations over voluminous examples."[2]
The genesis of Reddy's paper drew inspiration from work done by others in several disciplines, as well as linguistics.[3] Research on information theory had led Norbert Wiener to publish the seminal book on cybernetics, in which he had stated, "Society can only be understood through a study of the messages and communications facilities which belong to it."[4] Social-systems theorist Donald Schön examined the effects of metaphorical speech in matters of public-policy; he suggested that people's conflicting frames of reference were often to blame for communication breakdown.[5] Schön's frame-restructuring solution was similar in some ways to Thomas Kuhn's groundbreaking views on the shifting of scientific paradigms through what he called the "translation" process.[6] [7]
Research within linguistics (including the controversial Sapir-Whorf hypothesis[8] and Max Black's arguments against it),[9] coupled with Uriel Weinreich's assertion that "Language is its own metalanguage,"[10] prompted Reddy to approach the conduit metaphor's exposition and its possible impact on language and thought with caution.[11]
Reddy collected and studied examples of how English speakers talk about success or failure in communication. The overwhelming majority of what he calls core expressions involved dead metaphors selected from speakers' internal thoughts and feelings. Speakers then "put these thoughts into words" and listeners "take them out of the words." Since words are actually marks or sounds and do not literally have "insides," people talk about language largely in terms of metaphors.
Most English core expressions used in talking about communication assert that actual thoughts and feelings pass back and forth between people through the conduit of words. These core expressions and the few that do not qualify as conduit metaphors are listed in the paper's extensive appendix,[13] which itself has been cited by Andrew Ortony as "a major piece of work, providing linguistics with an unusual corpus, as well as substantiating Reddy's claims about the pervasiveness of the root metaphor."[14]
There are two distinct but similar frameworks in which the conduit metaphor appears. Four types of core expressions constitute the major framework. (In the following example sentences, the operative core expressions are italicized.)
These commonplace examples—
—are understood metaphorically. In 1., people do not actually "get across" concepts by talking; in 2., feelings do not really "come through to" people; and in 3., people do not in fact "give" to others their ideas, which are mental states. Listeners assemble from their own mental states a partial replica of the speakers'. These core expressions assert figuratively that language literally transfers people's mental contents to others.[15]
These examples—
—show that in 1., the speaker might be inexperienced in ensnaring meaning; in 2., be clumsy when putting it in; in 3., put it in the wrong place; in 4., compel words to accommodate meanings for which there is not enough room; in 5., fail to put in enough; or in 6., put in too much. These core expressions assert that speakers "insert" mental content into the "containers" represented by words with varying degrees of success.[16]
These examples indicate that sounds and marks can be "containers" for mental content:
These core expressions assert that words contain or do not contain mental content, depending on the success or failure of the insertion process.[17]
These examples—
—indicate that speakers and writers are responsible to a large extent for the mental content conveyed by language, and that listeners and readers play a more passive role. However, in 7., a reader can add something to the container that was not originally there. Overall, these core expressions assert that listeners must "extract" mental content from words.[18]
Instead of words, an "idea space" between people's heads can be the container for mental content. The conduit is no longer a sealed pipeline between people, but an open pipe allowing mental content to escape into, or enter from, this space. Three types of core expressions constitute the minor framework of the conduit metaphor.
These examples—
—show that speakers and writers can eject mental content into an external idea space outside people.[19]
These examples—
—indicate that mental content has a material existence in an idea space, existing outside people.[20]
The following examples—
—demonstrate that mental content from an idea space may or may not re-enter people.[21]
The italicized words in the above examples are interchangeable with a wide array of terms that label mental content, the containers in which the content may be placed, and the ways in which these containers may be transferred in the conduit-metaphor paradigm.
Reddy developed a logical apparatus for diagramming the conduit metaphor's many permutations in both frameworks. Mental contents (feelings, emotions, ideas, etc.) are represented by RM, which stands for "repertoire member." Containers (words, phrases, sentences, etc.) are represented by S, which stands for "signal." Thus, "I need to put each idea into phrases with care" can be rendered as the core expression put RM into S.[22] Reddy uses this logical apparatus throughout the appendix to his paper to clarify distinctions between metalingual expressions that use the conduit metaphor and the minority that do not.[23]
In order to examine the effects of the objectification of mental content in communication using the conduit metaphor, Reddy proposes an alternate, contrasting, "radical subjectivist" conception of communication called the toolmakers paradigm.[24]
A person's mental content is in fact isolated from others'. This isolation can be represented by a wheel-shaped compound, each wedge-shaped sector of which is an environment (a brain) bounded by two spokes and part of the circumference (right). They all contain differing amounts and types of plants, rocks, water, etc. (repertoire members). The wheel's hub has machinery that can deliver sheets of paper between sectors (communication). People use it to exchange crude blueprints (signals) for making tools, shelters, foods, etc., but they have no other contact whatsoever, and know of others' existence by inferences based on these blueprints.[25]
Living in a forested sector, Alex builds a wooden rake, draws three identical blueprints, and drops them in the slots for Bob, Curt and Don. Bob finds a piece of wood for the handle, but because he lives in a rocky sector, starts making a stone rake head. (Alex had not considered wood to be unavailable or wrong for the rake head, so it was not specified.)
When halfway done, Bob connects his stone head to the handle, realizes it will be heavy, and decides it must be a device for digging up rocks when clearing a field for planting. (He infers that Alex must be either very strong or has only small rocks in his sector.) Bob decides two large prongs will make his tool lighter, thus finishing with a two-bladed pickax. He makes three identical blueprints for his pickax and drops them in the slots for Alex, Curt and Don.
Alex assembles a kind of rock-pick, but must modify the design if a wooden, two-pronged head is to be strong enough. (He cannot see much use for the tool in his largely rock-free sector, sensing that Bob has misunderstood his rake.) Alex draws a second blueprint for the rake head and sends it out as before. Curt crafts a hoe for slicing cleanly through roots to clear out a swamp. Don creates a gaff for fishing.[26]
Blueprint users (language users) in the toolmakers paradigm can converge by inference on accurate replications of others' tools (mental content) after a laborious series of exchanges. Using the same diagram for the conduit-metaphor paradigm instead, the hub is a duplicator that can transfer actual materials and constructions among sectors, ending the isolation. No guesswork or construction is needed: Alex puts the rake in a special chamber, pushes a button, and precise replicas appear instantly in similar chambers for Bob, Curt and Don.
The subjectivist toolmakers paradigm embodies a language requiring real effort to overcome failures in communication, whereas the objectivist conduit-metaphor paradigm embodies one in which very little effort is needed for success.[27]
Although the toolmakers paradigm is available as a more accurate model of communication, the conduit metaphor is pervasive and difficult to avoid in English syntax and semantics. Thinking in terms of another model of communication is generally brief, isolated and fragmentary because of an entrenched system of opposing attitudes and assumptions.[28]
Reddy's tally of core expressions is about 140. Examining alternative ways of speaking about communication—either metaphorically opposed or neutral to the conduit-metaphor framework—results in a list of 30 to 40 expressions. Thus, 70% of the metalingual apparatus of the English language is based on the conduit metaphor. The influence of the remaining 30% is weakened by several factors.
Speaking carefully and attentively, it is possible to avoid conduit-metaphor expressions. For example, "Did you get anything out of that article?" might be replaced by, "Were you able to construct anything of interest on the basis of the assigned text?" Eschewing obvious conduit-metaphor expressions when communication is the topic is difficult. "Try to communicate more effectively" differs in impact from "You've got to learn how to put your thoughts into words." Reddy proceeds to show that even if avoidance were possible, it does not necessarily free people from the framework.[30]
A semantic pathology arises "whenever two or more incompatible senses capable of figuring meaningfully in the same context develop around the same name."[31] "I'm sorry" is an example of two contextually relevant meanings in collision. A person may expect an apology when the other wishes only to sympathize, or anticipate sympathy but hear an apology instead.[32]
Many other terms are ambiguous between mental content and the words "containing" it. For instance the word "poem" denotes a particular grouping of the sounds or marks (signals) exchanged between people. However, its use in sentences reveals that it can refer to thoughts or feelings (repertoire members). In this example—
The poem has four lines and forty words—"poem" refers to a text. The word-sense[33] can be labeled POEM1. However, in this example—
Eliot's poem is so utterly depressing—"poem" refers to the mental content assembled in its reading. The word-sense in this case can be labeled POEM2. Moreover, this example—
Her poem is so sloppy!—can be understood as either POEM1 or POEM2 (polysemy).
The ambiguity of "poem" is intimately related to the conduit metaphor. If words contain ideas, then POEM1 contains POEM2. When two entities are commonly found together, one of their names—usually the more concrete—will develop a new sense referring to the other (the process of metonymy). Just as ROSE1 (the blossom) developed ROSE2 (a shade of pinkish red) by metonymy, so POEM1 gave rise to POEM2.[34]
In the toolmakers paradigm, words do not contain ideas, so POEM1 cannot contain POEM2; therefore, a distinction between them must be preserved. Although there is only one POEM1 in most cases, the differences in mental content among people (and the difficult task of assembling it based on instructions in the text) mean that there are as many POEM2s as there are people. These internal POEM2s will only come to resemble one another after people expend effort comparing their mental content.
If language had been operating historically under the toolmakers paradigm, these two different concepts would not currently be accessed by the same word: talking about mental content and signals as if they were the same would have led to insoluble confusion. The ambiguity of "poem" would thus have been an incurable semantic pathology. However, the conduit metaphor can completely ignore it.[35]
"Poem" is a paradigm case for the entire class of English words denoting signals ("word," "phrase," "sentence," "essay," "novel," "speech," "text," etc.), demonstrating that semantic structures can be completely normal in one view of reality and pathological in another. This lends support to the theory that language and views about reality develop together.[36]
Evidence of the biasing power of the conduit metaphor can be found in fields outside of linguistics. Information theory, with its concept-free algorithms and computers as models, would seem to be immune from effects arising from semantic pathology, because the framework shares many attributes with the toolmakers paradigm. Nevertheless, there is evidence that use of the conduit metaphor has hampered investigators' attempts to develop the theory.[37]
Communication is the transfer of information (a selection from a set of alternatives). This set and a language (code) relating the alternatives to physical signals are established. A copy of each (an "a priori shared context") is placed with the sender and receiver. A sequence of the alternatives (the message) is selected for communication to the receiver, but the message itself is not sent. The selected alternatives are related by the code to energy patterns (the signal) that travel quickly and unmodified. Mathematics is used to measure quantitatively how much the received signal narrows down the possible selections from stored alternatives.
The similarity between the frameworks of information theory and the toolmakers paradigm is that
This analogy has withstood information theory's utility in simple, technical applications, but in biology, the social sciences, human language and behavior, it has been historically less successful. These attempts foundered by misunderstanding the conceptual framework of the theory rather than its mathematics. Reliance on ordinary language has made the information theory's insights less clear.[39]
The negative impact of ordinary language on information theory's use in other fields can be traced to terms the founders themselves used to label parts of their paradigm, telegraphy. The set of alternatives (repertoire members) were called the "alphabet." While true for telegraphy, Claude Shannon and Warren Weaver used it as a nomenclature referring to any set of alternative states, behaviors, etc.[40] The alphabet confuses the distinction between signals and repertoire members in human communication. Despite Weaver's particular interest in applying the theory to language, this fact went unrecognized.
Shannon and Weaver were also unaware that the choice of the term "message" to represent the selection of alternatives from the repertoire shared the same semantic pathology as "poem."
Because MESSAGE1 is the signal and MESSAGE2 is the repertoire members in communication, the reasoning is faulty. The ambiguity is trivial in the conduit-metaphor framework, but fatal for information theory, which is based on the idea that MESSAGE2 cannot be transmitted. Although Shannon and Weaver noted the distinction between "received" and "transmitted" signals based on possible distortion and noise, they wrote the word "message" on the receiving end of their paradigm.[41] [42]
Weaver employed many conduit-metaphor expressions; for example, "How precisely do the transmitted symbols convey the desired meaning?" [italics Reddy's].[43] He also contrasted two "messages, one of which is heavily loaded with meaning and the other of which is pure nonsense."[44] Weaver wrote as if repertoire members are sent, adding that the sender "changes the message into the signal" [italics Weaver's].[45] A code specifies how two systems relate, without changing anything; it preserves in the receiver the organizational pattern of the sender. Marks and sounds do not change into electrons, just as thoughts do not change into words. Shannon correctly wrote, "The receiver ordinarily performs the inverse operation of that done by the transmitter, reconstructing the message from the signal."[46] But conduit metaphors continue to appear in the form of "encode" and "decode," defined as putting the repertoire members into code and taking them out, respectively. In addition, because the theory conceives of information as the ability to copy an organization via nonrandom selections, the term "information content" is itself a misnomer: signals do something but cannot contain anything. The conduit metaphor has thus influenced the thinking of information theorists in a counterproductive way.[47]
Confusion between the message and the signal persisted for two decades as theorists in other fields of inquiry drew on the insights of information theory. Kenneth K. Sereno and C. David Mortensen wrote that "investigators have yet to establish a completely acceptable definition of communication".[48]
"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 Sereno & Mortensen's][49]Additionally, when they state, "The theory was concerned with the problem of defining the quantity of information contained in a message to be transmitted...,"[50] information is contained in a transmitted "message". If it refers to MESSAGE1, it is the conduit metaphor asserting that information is contained in the signals. If it is MESSAGE2, it is the repertoire members that are sent in signals, which contain measurable information. The insights of information theory have been challenged by using the conduit metaphor instead of the more accurate toolmakers paradigm, upon which its premises were initially based.[51]
The conduit-metaphor paradigm states that communication failure needs explanation, because success should be automatic: materials are naturally gathered, but misguided people expend energy scattering them. Conversely, the toolmakers paradigm states that partial miscommunication is inherent and can only be fixed by continuous effort and extensive verbal interaction: materials are gathered using energy or they will be naturally scattered.
Reddy explores some of the potential social and psychological effects of believing that communication is a "success without effort" system, whereas it is an "energy must be expended" system. The conduit metaphor objectifies meaning and influences people to talk and think about mental content as if it possessed an external, inter-subjective reality.[52]
Having discussed the conduit metaphor's impact on theorists within and outside of linguistics, Reddy speculates about its distorting potential in culture and society. He points out that
You'll find better ideas than that in the libraryis a conduit metaphor asserting that ideas are in words, which are on pages, which are in books, which are in libraries—with the result that "ideas are in libraries." The implication of this minor-framework core expression is that libraries full of books, tapes, photographs, videos and electronic media contain culture.
In the toolmakers-paradigm perspective, there are no ideas in the words; therefore, none in libraries. Instead, there are patterns of marks, bumps or magnetized particles capable of creating patterns of noise and light. Using these patterns as instructions, people can reconstruct mental content resembling that of those long gone. Since people in the past experienced a different world and used slightly different language instructions, a person unschooled in the language and lacking a full reservoir of mental content from which to draw, is unlikely to reconstruct a cultural heritage.
Because culture does not exist in books or libraries, it must be continually reconstructed in people's brains. Libraries preserve the opportunity to perform this reconstruction, but if language skills and the habit of reconstruction are not preserved, there will be no culture. Thus, Reddy asserts that the only way to preserve culture is to train people to "regrow" it in others.
He stresses that the difference of viewpoint between the conduit metaphor and the toolmakers paradigm is profound. Humanists—those traditionally charged with reconstructing culture and teaching others to reconstruct it—are increasingly rare. Reddy proposes that, despite a sophisticated system for mass communication, there is actually less communication; and moreover, that people are following a flawed manual. The conduit-metaphor influenced view is that the more signals created and preserved, the more ideas "transferred" and "stored." Society is thus often neglecting the human ability to reconstruct thought patterns based on signals. This ability atrophies when "extraction" is seen as a trivial process not requiring instruction past a rudimentary level.
Reddy concludes that the conduit metaphor may continue to have negative technological and social consequences: mass communications systems that largely ignore the internal, human systems responsible for the majority of the work in communicating. Because the logical framework of the conduit metaphor indicates people think in terms of "capturing ideas in words"—despite there being no ideas "within" the ever-increasing stream of words—a burgeoning public may be less culturally informed than expected.[53]
Since the publication of Reddy's paper in 1979, it has garnered a large number of citations in linguistics, as well as a wide spectrum of other fields of inquiry. In 2007, a search at Web of Science [54] revealed 354 citations broken down roughly as follows: 137 in linguistics; 45 in information science; 43 in psychology; 38 in education; 17 in sociology; 15 in anthropology; 10 in law; 9 in business/economics; 8 in neurology; 7 in medicine; 5 in political science; 4 each in the arts, biology, environmental science, and mathematics; and 1 each in architecture, geography, parapsychology and robotics.