Mixcd metaphor of'thc intos-mation superhighv~av as ocean in which a fishing net circciges Up infomiation. Interactive multimedia interface to an information network by 7.~1lan Nemes Nemeth.

7

1

i

Metaphors and Machines

Metaphor, Bein$ and Computer Systems Design

The effective use of metaphor seems to be an important consider ation in computer interface design. Many computer systems incorpo rate pictures, ideograms, and icons that depict objects, tools, files, and other operational and organizational devices displayed on the computer screen, and the use of the hand-held "mouse" or touch screen as a means of pointing at, and clicking on, objects. The com puter screen is arrayed with metaphorical objects that exhibit certain properties and with which one can interact. According to one sys tems developer, Erickson,~ "Just about everyone at Apple knows the phrase 'desktop metaphor' and fervently believes' that a good meta phor is essential to an easy-to-use human interface."2 Of coul-se, the use of metaphor in computer systems design goes deeper than thc simple deployment of pictures and familiar key words. The struc tures of the operating system3 and other computer programs and subroutines are commonly designed as a network of (invisible) ob jects. These objects have properties, generate and receive "events," and send "messages" to one another.4 They take on meanings of the software designer's own choosing, ranging from the abstract (sucl1 as the algebraic variable x) to the more concrete, though bizarre, "demons" or "autonomous agents" that monitor what is going on in the computer system and act in response to some problem. Com puter hardware can also be considered in metaphorical tenlls. Manv electronic components are named after familiar entities: memory, gate, chip, processor, and so on. In these senses, the entire computer

system is imbued with metaphor, from the structure and configuration of hardware to icons and the designation of objectlike names attached to program subroutines.

Adopting metaphor as a design concern seems to appeal to pragmatically oriented systems designers. It suggests an engagement with the world of the computer user and programmer through familiar and recognizable objects, as opposed to esoteric commands and formal logic. An emphasis on metaphor also represents a liberal attitude to design. Through such emphasis, design is cast largely in terms of devising appropriate metaphors rather than solving a problem through theoretical analysis. Design becomes a process of comparing the efficacy of metaphors rather than matching solutions to problems through objective criteria. A metaphorical approach to design can be cast in much looser and more pragmatic terms than the idea of method suggests. Metaphor also elevates the role of the imagination in design. The screen is an imaginary sheet of blank paper; a screen window is an imaginary opening into a world of information. Metaphor also implicates the human body. In pointing and clicking on the computer screen, we imagine we are touching and grasping real objects, a precursor to "virtual-reality" systems and "tactile computers."

There are further ways in which metaphor appeals to computer systems designers. (Traditionally, the study of metaphor belongs within the study of rhetoric, language, literature, psychology, and philosopl1y.) The appeal may reside in the implication that a comouter system is a blank slate or a Itlmp of clay for designers to manipulate in any way they choose.5 We can fashion whatever we want Otlt of the computer_a desktop, a filing cabinet, a stage play_much as we can make almost anything Otlt of words. In other design areas, such as architecture or mechanical engineering, the designer feels more constrained by tools, techniques, materials, and a physical setting. It seems that principle and method assume greater importance. By way of contrast, the medium of the computer-software designer is pure, malleable strings of electrical impulses, subject to laws, but waiting to be fashioned through whatever metaphors we desire.6

Not everyone agrees that talk of metaphor is useful. Some systems developers operate within the pragmatic mode of design but are

1

1

l

skeptical of metaphor. Nelson,7 for example, maintains that the "metaphor business has gone too far."S He al-grlles that the linl; hltween the objects on the computer screen (the pictul-e of a trash can) and real objects (a real trash can) is very tenuous, and it hinders rather than helps human-computer interaction. Such pictures are mnemonic gimmicks. They assist memory but do little more, and, according to Nelson, the metaphors are usually ill chosen. They become a dead weight_the whole computer program has to be structured around them. The metaphors become "forced." They also introduce inconsistencies into the operation of the program_for example, the trash can is used to dispose of files but also to eject disks (ready for safe-keeping).

What is the alternative? Nelson proposes a consideration of the whole of a system rather than a collation of assorted metaphors: "The alternative to metaphorics is the construction of well-thought-out urlifying ideas, embodied in richer graphic expressions that are not chained to silly comparisons. These will be found by overall virtuality design . . ., and not by metaphors, which I consider to be using old half-ideas as crutches."9 Nelson's criticism appears to be directed mainly against a particular kind of metaphor use, the fairly "trivial" one of having a pictograph Ol1 the screen that looks like a folder Ola trash can and expecting it to behave like one. Nelson advocates that systems developers follow the pattern set by film production. A film is planned and directed to form an artistic whole. But of course, we can regard what Nelson is proposing here simply as a different metaphor, a grand, holistic metaphor of computer software as a movie.

Nelson points out that some computer systems have had this holistic "nonmetaphorical" character. Apparently the Visicalc spreadsheet had no reference to anything that had gone before: "To replicate a column and its formulas corresponds to nothing that was on earth previously; and when metaphoric thinking was dismissed, it could be designed cleanly with no reference to anything that harl come before." ~_ Nelson describes this as discovering a new principle. But then it could equally be described as discovering a new metaphor, and this by the melding of existing metaphors: work sheets, bookkeeping, timetables, truth tables, formulas, and so on. Thc

sound advice Nelson is offering can be translated simply into metaphorical terms: do not get fixated on trivial and obvious metaphors, allow metaphors to play against one another and new metaphors to emerge, situate yourself in the problem domain rather than immediately fixating on a particular metaphor, look for the grand metaphor, do not take metaphors too literally by fixating on one-to-one correspondences between the metaphor and the object, and carefully evaluate your metaphors.

Kay also criticizes metaphor:" "My main complaint is that metaphor is a poor metaphor for what needs to be done."'2 He posits another term, "user illusion," contending that it may be helpful for part of the computer screen to resemble a sheet of paper onto which one can type words or draw, but it is the "magical" qualities of that paper, those that escape the analogy, that are the most valuable. What can you accomplish with magic paper? You can layer things on it, animate drawings, and erase much more easily than on real paper. If computer systems designers simply followed the metaphor, the screen paper would be no more enabling than real paper. To further illustrate, Kay draws out some of the shortcomings of HyperCard, a high-level computerized database system that exploits the metaphors of stacks, cards, buttons, and fields. According to Kay, the difficulty is that you can write only on cards, not on buttons; you can place cards in stacks but not stacks in cards, and so on. Kay argues that the distinctions between stacks, cards, buttons, and fields are unnecessary; a much better idea is simply to have random containers. Kay's advice is that rather than have arbitrary and limiting metaphorical distinctions, the systems designer should adopt a more unitary concept. But here again the appeal is to more efficacious metaphors, random containers rather than differentiated objects, holistic metaphors rather than reductive and disparate ones. As we will see, it is precisely the "magical" qualities of metaphors that enable them to work. Every metaphor brings to light differences, and a substantial part of the power of metaphor resides in difference.

But the application of metaphor to computer systems extends even further than is suggested by the discussion so far. As most sys

tems developers and commentators acknowledge, the appeal to metaphors extends beyond what appears on the computer screen or in the computer program. Our whole conception of the computer is driven by a varied range of "metaphorical orientations." These orientations are apparent when we think of the common terms "interface, " " memory, " "art)ficial intelligence, " "problem solving, " "state space," as well as many specialized terms, such as "motherboard," "bandwidth," "register," "packet switching" and "system crash." Each of these terms can be regarded as a metaphor that associates a computer component with some other domain or human practice. Such metaphors are often anthropomorphic: "memory," "intelligence," and so on. So the metaphorical orientation is of the computer as a human entity. In arguing for a new view of the computer and how it might fit into the workplace, commentators frequently appeal to some new metaphor or metaphorical orientation: Laurel promotes the idea of the computer as theater; Turkle as Rorschach inkblot; Kay as medium; Nelson as movie machine; Weiser as ubiquitous facility_pens, paper, and the electricity grid. The computer itself also serves as a metaphor for other things. The computer is implicated in various conceptions of the workplace, education, society and the workings of language, and the mind and communication. A prime example is the famous information-processing metaphor of cognition. So metaphor practically permeates our entire conception of the computer.

Several major questions underlie the issue of metaphor in computer systems design. Is it possible to get back to some reality beyond metaphor? Are all metaphors feasible in all situations? Are there any constraints posed by the computer medium that restrict the possibility of certain metaphors? Is computer systems design metaphorical all the way through? Is metaphor so ubiquitous and universal as to be a meaningless concept? What of the grand metaphors through which we see ourselves in the computer? In what follows, I consider the different views on metaphor and how different traditions of~ thinking impinge on the issue. In the process, I develop some valuable insights on the efficacy of metaphor in computer systems design. We will also be in a position to evaluate the role of the design models outlined in chapter 6.

What is metaphor?'3 According to Aristotle, "metaphor consists in giving the thing a name that belongs to something else."'4 Metaphor is an example of imitation (mimesis), which, according to Aristotle, is the integrating principle of poetry, tragedy, comedy, and music. Aristotle also shows how metaphor is an integral part of "style" in rhetoric.'5 According to Aristotle, metaphor is also the use of figurative, exotic, or ornamental language. The appeal of figurative language is that "the discourse must be made to sound exotic; for men are admirers of what is distant, and what is admired is pleasant."'6

Terms related to metaphor include trope, which is a general term for the figurative use of a word (for example, to describe someone as radiant, normally a property of a hot or luminous object); synecdoelle, in which we substitute a part of something for the whole (as in one hundred head of cattle); metonymy, in which we substitute an attribute for the thing that is meant (as in the use of "the crown" to refer to "the monarch"); catachre.sis, in which we misuse, deliberately or otherwise, a word that sounds similar; and ellipsis, in which we deliberately omit words in a sentence. As we will see, the most important terms related to metaphor are analo~, and simile. Derrida explores these and other terms in an illuminating article on metaphor, "White mythology."'7 Derrida is one of the most celebrated practitioners of the use, and even deliberate abuse, of such devices in philosophical argument.

For Aristotle and mtlch of the philosophical tradition up until the twentieth century, metaphor was an element of language to be used cautiously. Metaphor was regarded as an embroidering of truth. It was Nietzsche who vividly drew attention to the impossibility of formulating truths independently of metaphor.

What therefore is truth? A mobile army of metaphors, metonymies, anttlropomorphisms; fi1 shol-t a sum of human relations whicl1 became poetically and rhetorically intellsified, metamorpllosed, adorned, and after long u.sage seem fixed, canollic ancl binding; tmths are ilhlsiolls of which one has forgotten that they are ilklsions; worn out metapllcJrs whicl1 have bc-come powerless to afJ~c ct the sense, COillS \VhiCIl have their obvel-se effacecl and now are 110 IcJnger of account as coins but merely as metal.~8

1

i

C:oupled with metaphor's rise in importance in the twentieth century is an appreciation, within the contemporary German ancl Frellcll philosophical tradition at least, of the importance of rhetol-ic and all its devices. As indicated by Aristotle's treatise on the subject, metaphor comes within the study of rhetoric. The primacy of rhetoric over logic is an important theme in postmodern writing. According to Gadamer,'9 the ubiquity of rhetoric is "unlimited." Even the pursuit of scientific understanding comes within its ambit: "There can be no doubt . . . about the fundamental functioll of rhetoric within social life. But one may go further, in view of the ubiqtlity of rhetoric, to defend the primordial claims of rhetoric over against modern science, remembering that all science that would wish to be of practical usefulness at all is dependent on it."2" In the study of rhetoric, we consider the means by which we persuade, including conventions, systems of legitimation, the authority of the speaker (or writer), the authority of sources, the nature of the community in which the discourse is taking place, the situation of the writer and the reader, and the metaphors used. According to Ricocur,2, the study of rhetoric has been malignecl through sopllistry, but also through the rise in prominence of logic, grammar allcl class)fication: "[R]hetoric is given over to playing with distincticnls and class)fications. The genius of taxonomy occupies the space deserted by the philosophy of rhetoric."22 So the study of metapholfalls within this tradition of the study of rhetoric, and in appealing to metaphor we appeal to a tradition of thought dating back to anticlui0'. The tradition has undergolle a transformation. Now rhetc~lic and metaphor are regarded by many contemporary thiltkers 110t merely as embroidery but as essential to truth and understallditlg.

How does metaphor work? To speak metaphorically is simpl.v to relate two entities (or terms) through the verb "to be" (or the c opula "is")_a house is a machine, that person is a beast, design is following a procedure_or the preposition "as"_I regard tllc hotlse as a machine, men as animals, design as following a plOCC'dure. The juxtaposition of the terms in the metaphor commolllN (but not always) involve assigning an instance to a class, or a spec ic s to a genus, thotlgh obvious metaphors (such as a house is a maclline)

usually have the appearance of assigning an instance to the "wrong" class, or a species to the "wrong" genus.

Metaphor is not regarded only as a linguistic phenomenon. For such writers as Goodman,23 metaphor is implicated in perception: I see the drawing as a square and a circle overlapping, the floor plan as a flow diagram, the arrangement of pixels on the screen as a sheet of paper. "Seeing as" is a basic phenomenon of perception_so too with sound. According to phenomenologists, we do not hear abstract noises, which we then interpret, but we hear sounds immediately as car engines, as bird calls, as people speaking. According to this view, we are constantly engaged in metaphorical projections. We project one term, concept, or situation onto another.

As a further means to understanding metaphor, it is helpful to look at the view of metaphor generally attributed to Aristotle. This is the substitution, or comparison, view of metaphor. The substitutic>n view asserts simply that it is always possible to rephrase a metaphor in literal language. To speak metaphorically is to embroider plain language. So to say that a house is a machine is to imply that a house has certain properties: it has simple functionality, it has inputs and OtltpUtS, it can be mass produced, and so on. According to this view, a metaphor is an abbreviated form of simile. A simile has the form A is like B. The statement could be expanded to "a house is like a machine," and then we may append a list of properties: "in that it has certain properties." According to the comparison view, when we extract meaning from the statement "a house is a machine," we invoke a list of properties pertaining to a machine and compare them with those of a hotlse. The substitution view also treats metaphors as implied analogies. An analogy has the general form A is to B as B is to C_the supply of services is to a house as energy is to a machine.

Seen in this light, some view metaphor as not belonging to serious discourse. Metaphor obviously features in poetry but not in scientific or philosophical study, or in other cases when we wish to make our meaning absolutely clear. The rationalist, empiricist, and Enlightenment projects were bent on cutting through the obscuring effects of figurative language. But this concern was also revived by logical positivists in the 1930s who maintained that figurative language has

meaning only insofar as it can be translated into literal language. Furthermore, logical positivists thought that confilsion withill figurative language was responsible for many philosophical and logical errors.24

This concern to get back to the realities of a situation unel1ctlulbered by metaphor echoes the argument outlined by Flesse about the role of models in science.2'' (As we will see, a model is arguably the formal treatment of a metaphor.) There are those who say that science does not need models_such as the wave and particle models of light_but simply sound theories, ruses, and formulas that unite observable and measurable variables and predict outcomes. A model is merely a pedagogical tool we can dispense with once we have grasped the principles. Certainly there is more than a hint of this suspicion of metaphor in the criticisms leveled at computer interface design. In analyzing the metaphor of the computer screen as a sheet of paper, we think of the properties of paper and those of the computer screen. A sheet of paper is flat, can be written Ol1, can be overlaid with other sheets, and so on. The screen display is also flat, can be written on and overlaid, but it is ustlally vertical rather than horizontal. You cannot fold it down the middle or make it into a paper airplane, and you can do lots of things with it that are impossible with real paper. Seen in this light, the paper metaphor appears unnecessary, except perhaps as a temporary learnillg device for novices. The substitution view of metaphor suggests that the paper metapl1or can be substittlted by a literal collection of concepts. We could strip away the interface metaphors and get down to the realities of the system, which should be founded on SOUllCt principles.

But some metaphors clearly cannot be reduced to lists of features. Poetry provides many obvious examples of this. Some poetical cxpressiol1s appear to lose most of their meaning if they are paraphrased in literal terms. To analyze "Juliet is the sun" in the terms outlined above (as a list of features) wotlld involve an endless regress of metaphors. As a simile, Juliet is like the sun in that they botll radiate warmth, are spectacular, give life, rule over the day, ancl so on, but Juliet and the Still do not exhibit these features in thc same literal sense. Each feature is also a metaphor that needs to bc

elaborated further. Furthermore, the expression is a poetical one, whose value is considerably diminished by such analysis. The metaphor is irreducible. The metaphor is explicable rather in terms of understanding Jtlliet in a new way by associating her with the sun. The two terms of the metaphor interact to produce a new and extended meaning to "Juliet." This new meaning is clearly open to wide interpretation and will vary depending on context and what we understand by the two terms.

Counter to the substitution view of metaphor, the interactionist view asserts that metaphors are always of this kind, whether poetical or not.26 According to Black, a metaphor has a primary and a secondary subject. In the metaphor "a house is a machine," the primary subject is "house" and the secondary subject is "machine." The secondary subject is to be regarded as a system rather than an individual thing. So "machine" is a set of concepts rather than an instance of a machine. It is clearly not that machine over there_the old electrical fan in the corner, or any other particular machine. On the other hand, "house" can be a particular house or, in this case, any house. A metaphor commonly ascribes an instance to a class or a species to a genus. According to the interactionist theory, the metaphor works by projecting a set of "associated implications" onto the primary subject_machineness and its entailments are projected onto the concept of a house. The two subjects interact in several ways. The primory subject incites the hearer to select some of the secondary subject's properties: being functional, having inputs and OUtptltS, ancl so on. The metaphor constructs a parallel "implication complex" that can fit the primary subject: a house is functional, has inputs and OUtptltS, and so on. There are also parallel changes in the secondary subject_we regard in a new light what it is to be a machine. By virtue of this usage a machine takes on hotlselike qualities. New metaphol-ic uses for "machine" are opened up as well as new ways of conceiving of machines.

Althotlgh Black highlights the notion of properties, this is not to reduce metaphors to lists, as in the substitution view. The interactionist view holds that any analysis of a metaphor in terms of feature or property lists is a contextual actisity, an expedient for particular kinds of analysis. The workings of any particular metaphor can never

be fully captured in an "objective" sense. For example, to regard part of the computer screen as a sheet of paper (the paper metaphor) cannot be reduced to lists of features. The juxtaposed terms enhance understanding in some way, and that understanding is C011tingent on the situation. To understand a metaphor is always to interpret it, and there are different interpretations according to context. The computer systems designer exploring the metaphor may focus on writing conventions, the background/foreground property, rectangular format, orthogonality, and erasure. There is no exhausting such lists, and in certain contexts, analysis might also focus on other, less-obvious properties, such as reflectance, color, stiffness, topology, and so on. But a metaphor need not be broken down in order to be useful or meaningful, and our everyday, unreflective engagement with a metaphor does not require such analysis.

Much contemporary debate about metaphor elaborates these two basic understandings: the substitution view (that every metaphor has a literal equivalent) and the interactionist view (that metaphors are irreducible and rely on the context-dependent interaction between two terms). Clearly art)ficial-intelligence research and studies into cognitive modeling find greatest potential in some variation of the substitution view, because it implies that metaphors can be represented and controlled.2'

There are two other poles in the debate about metaphor. One position holds that the designation "metaphor" is simply a category of utterance (like a question or a command), and metaphor is not a fundamental constituent of langtlage. Accorcling to this view, all language is literal, inchlding such statements as "a house is a machine" and "Juliet is the sun." The seconcl view maintaills that all language is metaphorical, including such matter-of-fact utterances as "that is a table" and "the earth revolves around the sun."

The foremost proponent of the literal langll.lge view is Daviclson.2 He maintains, "Metaphors mean what tile wolcl.s in their most literal intel~pl-etatioll. meall, and notlling more."2" B`Zt Daviclson does not thereby hold to the substittltioll view. Metaphors are not re~t''eible to

literal language; they are literal language. Davidson contends that there are problems with the interactionist view. The interaction view of metaphor regards metaphors as containers and suggests that they have a special cognitive status. According to Davidson, metaphors do not convey ideas; a metaphor does not have a special meaning. As a language theorist influenced by pragmatism (along with Austin, Searle, Wittgenstein, and most of the proponents of the interactionist _iew), Davidson argues that "metaphor belongs exclusively to the domain of use":30 "If we are to think of words in metaphors as directly going about their business of applying to what they properly do apply to, there is no difference between metaphor and the introduction of a new term into our vocabulary."3' Davidson furnishes a colorful illustration. He gives the example of teaching an alien from another planet the concept of "floor." We may teach the concept by pointing to the floor and presenting contrived pedagogical situations in which the floor features in some use context, much as Robillson Crusoe might have taught English to his servant Friday. Then we accompany the alien into outer space to visit its planet. En rotlte, we pOillt through one of the portholes to the spectacle of the rececling earth and exclaim "floor!" We would probably assume that to refer to the earth in this context as the "floor" is a metaphor. But according to Davidson, whether the declaration is meant metaphorically or not makes no difference to the alien. Nor does it matter. As far as the alien is concerned, the term "floor" is simply given a new context of use. Who is to say where the literal meaning of "floor" ends and its "metaphorical" use takes over? Davidson is careful to point Otlt that the new use of "floor" is not simply an extension of a common use. Words are constantly being used and reused in different contexts. The quest for original uses will not lead IIS vel:~ far.

In view of this, we could say that to designate a part of a colnpllter screen as a sheet of paper works by virtue of a particular speech community's acceptance and use of the term in that context, rather than some special cognitive operations associated with the special notion of metaphor. In fact, some terms appear to have greater currency in the new context. How often do we use the term "desktop" other than as a descriptor for a computer screen?

To emphasize how words work in a context, Davidson also casts the issue of metaphor in terms of the nature of a lie_some commentators prefer the terms "myth" or "fiction." According to Davidson, the major difference between simile and metaphor is that "all similes are true and most metaphors are false."32 (We will forego considering poststructuralist revisions of notions of truth.) Only when we take a sentence to be false do we entertain its possibility as a metaphor. Then we "start to hunt Otlt the hidden implications."33 How are all similes true? Because everything can be deemed to be similar to everytl1ing else in some sense, all similes are trivially true. A computer is like anything we care to imagine, in some sense_a tool, a medium, a person, a book, even an umbrella or a steam train. On the other hand, most metaphors are statements that are patently false, or fictional. "A house is a machine" is false; so too "the computer screen is a sheet of paper" and "a computer is an intelligent assistant." Of course sometimes metaphors are statements that are true but in a very trivial sense. Such expressions as "business is business" and "no man is an island" are so trivially true that we are prompted to take the expressions metaphorically. We can therefore say that except for the trivial case metaphors are untrue statements that, by virtue of their context, are not to be dismissed but are worthy of consideration and have particular uses. According to Davidson, metaphors are untrue statements that are not lies.

How do we decide that a statement is a lie? Whether or not a statement is a lie or a metaphor depends on context. Davidson gi_es an example. The statement "my neighbor is a witch" could be one of three things. If we believe in witches, it could be either a lie or a truth. Or it could have a metaphorical meaning. (Or it could be a mistake.) The final arbitrator as to which meaning applies is simply the context of the utterance.

What makes the difference between a lie and a metaphor is not a difference in the words used or what they mean (in any strict sense of meandlg) but in how the words are used.... What distinguishes metaphor is not meaning but use_in this it is like assertion, hinting, Iying, promising, or criticizing And the special use to which we put language in metaphor is not_cannot be_to 'say something' special, no matter how indirectly. For a metaphcJr says only what shows on its face_usually patent falsehood or an absul-ci

trutll. And this plah1 truth ol falsehood needs 110 paraphrase_it is given h1 the literal meaning of the words.'4

According to Davidson, there is no content to be captured by a metaphor; rather the metaphor makes us notice certain things.35 (Davidson's critics deny his charge that they see metaphor as having "content" in the simple way he suggests.) Davidson's view of metaphor is considerably closer to the interactionist view than it is to the substitution view. Davidson effectively conflates the phenomenon of lallgtlage into a single pragmatic system of words in context. This word usage does not require that words be broken down into definitiOIIS, schemes, or lists of similarities, and we do not need a special theory to account for metaphor.36 The counterview to Davidson (that all language is literal) is that all language is metaphorical. I will return to this proposition in the next section.

Two additional polarities feature in discourse on metaphor. These cast filrther light on the implications of Davidson's view. Nearly all comlllentators on metaphor offer some account of the distinction hetween living and dead metaphors. For some commentators, the tc rm dead metaphor is synonymous with "the literal." A living metaphor is a juxtaposition of terms that still has some power to strike us as incongruous, shocking, novel, strange, interesting, or in other ways informative_a house as a machine, Juliet as the sun, a computer as an intelligent being, the computer as theater, and so on. In Davidson's terms, these expressions are metaphors because we cannot account for them as lies. On the other hand, a dead metaphor is an expression that no longer invokes investigation, reflectiOII, or new insights. Examples of dead metaphors include "this object is a table," "a computer is a calculating machine," "a pixel is a colored dot on a computer screen." There are many less-obvious examples. Ascribing memory to a computer chip could be regarded as a dead metaphor; so too are computer file, command, menu, and so on. These terms are so much a part of computer parlance and practice that they are generally taken "literally." We have mostly forgotten that the terms are "borrowed" from other fields_memory is a humall faculty, a file is a folder or a box to store documents, a command is an order, a menu is a list of food dishes.

But who is to say that memory is not simply the storage of information, a faculty shared by both people and computers; that a file is not simply a means of storing doctlments, whether in cardhoard or magnetic media; that documents are not simply information, whether in paper or electronic form; that a command is not simply an order, whether to a person or a machine; or that a menu is not simply a short list of items for selection, whether presented in a restaurant or on a computer screen? What Davidson's arguments brin~s; to light is that all terms are borrowed in some way or other. The distinction between living and dead metaphors follows the distinction between the metaphorical and the literal. The distinction is determined by use.

A full account of the debate surrounding Davidson's ideas is beyond the scope of this book. As I will show, Heideggerian concepts of truth as disclosure rather than correspondence provide a fruitful account of language and metaphor use contrary to Davidson's view.

Furthermore, Davidson's theories do not adequately account for the privilege that certain words carry from one context of use to another. It is not that Davidson is wrong but that the literal language view seems to close off a whole domain of fruitftll inquiry into thc privilegillg of terms. The domain of computerization pro\ides a goocl illustration of privileging. Ascribing human terms to the C0111puter and computational terms to people are not matters of inclif ference or coincidence. They indicate, among other things, a privileging of certain conceptions of ourselves and the world tllat has a history. If we follow the arguments of Heideggcr and Cl-itical theorists, then linking the computer to human intelligellce pt-Ovides one of many examples of our technological "enframing." Pre- Socratic concepts of thought have given way to the technological. The discourse on metaphor provides a useftll forum fol analyzing these aspects of such terms as "computer" and "intelligence." Rather than focus on incli\idtlal terms, wc can illSpCCt tile "haggage," or "entailments," that certain metaphors carry, their "structtlres" and relationships, and how they have risen to prollli

nence.

The counterview to Davidson is that all language is metaphorical. LakoffandJohnson are among the foremost proponents of this view. They also offer an account of how we privilege certain metaphors.

Lakoff and Johnson argue that the key to this privileging lies in grounding language in the human body.37 How does this work? Metaphors are related to one another in ways that we can analyze as structures_metaphor structures. According to Lakoff andJohnson, our direct, early, and shared experiences dictate and shape basic metaphor structures, which we call on when going about our dayto-day business. Lakoff cites evidence from linguistics, particularly cross-cultural studies, that abstract ideas, such as those pertaining to how we describe thought, are couched in terms that belong to the world as we interact with it physically. This is particularly evident in the use of categories. The category names to which we have immediate affinity, such as "table," "chair," "tree," and "dog" (basic categories), are more important to us than the more specific categories of "writing table," "Breuer chair," "elm" and "cocker spaniel."38 The argument is that it is more important to distinguish between a table and a chair than between different types of tables, say a dining table and a writing table, because of the way we use the objects. The basic categories are also more useful in day-to-day living than the more generic terms "furniture," "vegetation," and "animal." Lakoff also notes that because of their widely different bel~aviors, it is vitally important to make certain distinctions among the animal class. We are not generally content merely to group dogs, cats, and spiders together in the category "animal." Beause of our involvement in variotls pleasant and unpleasant experiences, we are more likely to declare "there is a spider in the bath" than "there is an animal in the bath." So our whole understanding of what constitutes an object and how we designate it relates to our communicated experience.

A similar analysis can be presented for our understanding of "abstract" objects. A similar hierarchization of categories applies to action words, such as "seeing," "touching," "running," and

265 Metaphors and Machines

"making." These verbs denote basic category terms that are experientially based. Of course, all category terms vary as our experience changes and as our interests and vocations develop. They also _ar\; across cultures. Language, however, has a conserving effect on these basic categories.

Even language pertaining to more complex aspects of our experience retains the vestige of these basic experiential terms_seeing, hearing, touching, and making. According to Lakoff and Johnson, abstract and complex activities, such as thinking and designing, are understood and described in these basic category terms. So we see the answer to a problem, are deeply touched by a situation, and make up our minds.

The next step in Lakoff and Johnson's argument is to see that categories of objects and actions do not exist merely in isolation but are formed into experiential gestalts_basic metaphor structures. Lakoff identifies several of these. There is a metaphor structure pertaining to containment: "a schema consisting of a boundary distinguishing an interior from an exterior. The container scheme defines the most basic distinction between in and out. We understand Otil own bodies as containers_perhaps the most basic things we do are ingest and excrete, take air into our lungs and breathe it out."3" Other metaphor structures pertain to paths, links, forces, balance, the up-down orientation, the part-whole relationship, and the cen~~r r`-rinl~f~rv rc- h~tionshiD.4~, Similarly, a journey schema is derivec! from basic experiences. Tied in with the journey metaphor are notions of intention and means. In order to satisfy a particular desire, a small child may need to embark on a journey across the room. According to Lakoff, this experientially based metaphor structure incorporates and sustains our basic ideas about intention and causality. Lakoff end Johnson use these basic and simple metaphor structures in analyzing more-complex linguistic and social phenomena. So exchanges within relationships between people can be seen as driven by certain dominant metaphors (for example, dialogue as war). These can be further broken down to particular schemes Ol basic metaphor structures.

According to Lakoff andJohnson, metaphor structures have their own internal relationships. How is it that certain aspects of a meta

phor structure gain primacy over other aspects? Lakoff points to the importance of the phenomenon of metonymy: "It is extremely common for people to take one well-understood or easy-to-perceive aspect of something and use it to stand either for the thing as a whole or for some other aspect or part of it."4i In the case of a journey, we may take the origin to stand for the whole thing. In response to the question "How did you get here?" the answer "I got on a bus" is often aufficient. The answer does not require a long explanation, because the rest follows from our knowledge of the bus- ride experiential gestalt. Of course, using another part of the journey metaphor, as in the answer "I got off a bus," is less usual and may lead to confusion. We may add that some computer screen icons appear to work metonymically (as do logos, symbols, crests, and so on). They depict a part of something in order to draw our attention to the whole_a pencil tip to depict a pencil, the letter A to indicate text, a compass to depict a drawing system, and so on. Metaphors have these and other internal structures.

According to Lakoff and Johnson, our understanding of logic is also driven by metaphor. They identify three main metaphors. The first is that of the experience of containment and its transitivity. The statement "all As are Bs" followed by "C is in A" informs us that therefore C is in B. This is a characterization of the syllogism. The metaphor of containment comes through our experience of inside and outside, supported by basic bodily experience. The containment schema is evident in everyday expressions, such as "let out your anger," and in comments about supposedly abstract reason, such as "let's start out from the following assumption" and.'that assumption will lead you astray." A second metaphor is that of force. The conchlsion from a set of logical propositions is considered to be inevitable. We are led to the conclusion through the force of inference. So we get another characterization of the same syllogism: A implies B, and A is true; therefore B is true. Force as the bodily activity of pushing against something leads to ideas about inevitability and causality. A third metaphor is that of balance. We weigh evidence; in makillg logical judgments, we balance conflicting requirements; and, in mathematical reasoning and formal logic, we balance both sides of an equation or a logical clausal statement. Schon also elabo

267 Metaphors and Machhles

rates on this metaphorical theme of balance.42 The use of this metaphor is evident where the consequences of certain actions have to be weighed against each other. We talk of the weight of evidence being in favor of a particular decision. Ethical problems are frequently discussed in these terms_"on balanceit seemed appropriate to favor the needs of those dependent on public transport."

Schon adds that we make liberal use of such metaphors as working with tools ("sharpening our wits") and social interaction ("my will and reason are at odds") in understanding thought. There are also metaphors of social processes. Here there are different parts to the self: "I told myself . . ." "my conscience told me . . ." "my will and reason are at odds." Under this metaphor structure, decision making (or designing) becomes a conversation between "internal advocates" of different kinds of rationality (or advocates making use ot different metaphor structures)_"It seemed logical to keep the structure simple, but I wanted the design to be a bold statement." There are also metaphors of mechanism and dynamism in which the mind is a machine. According to some process models, decisions appear as switches or taps. Schon suggests that arguments abotlt freedom and determinism are also essentially appealing to the mechanistic metaphor structtlre.43

How does the hoclily basis of lallgtlage and metaphor impinge on computer systems? There are two major ways. First, Lakoff andJohllSOII'S arguments linking reason to bodily experience, in keeping with the tenets of pragmatism, defuse the claim that reason is a ptll C' and transcendent phenomenon. In so doing, they challellge thcclaim that the computer may someday be a repository of reasoll, an intelligent device.44 They also challenge many of the models of chapter 6. Human cognitive ability does not reside in propositional logic but in such phenomena as metaphor use, which are irrcclllcible to propositions. As long as cognitive science and artificialintelligence research focuses on propositional logic, it will never meet its more ambitious goals. Second, Lakoff and 1ohnson's arguments indicate that the computer is not an abstract "reasoning" device, whose intemal logical structure renders it indepelldent of human concerns. In the same way that Johnson clemonstrates the presence of the body in the mind, we can readily point to the pres

ence of the body in the computer, or in any other human artifact. As confirmed by many critical commentators, the computer is not an impartial, value- neutral tool to be used for good or ill. But, at a more pragmatic level, Lakoff and Johnson also indicate that the computer is an embodiment of practical, bodily based concerns.

Following I,akoff andJohnson what is the bodily basis of the computer? We need look no further than the prevalence of the containment, force, and balance schemes in the conception and design of computer systems, or indeed of any machine, for evidence of the presence of the body. As far as containment is concerned, the registers in the microprocessor contain values, files contain data, and directories are organized hierarchically as nested containers. As far as the metaphor of force is concerned, the computer clock ticks away indefatigably at thirty thousand or more pulses per second, passing binary strings into and out of registers and logic circuits to ensure that the logical instnlctions produce their effects on the data. On the subject of balance, logic circuitry ensures that a statement like "A" balances with a statement like "not not A." The basis of the modern-day computer in binary logic also fi0s Lakoff andJohnson's account of the primacy in human cognition accorded to the metaphor structure of containment: in/out, near/far, on/off, right/left, up/down. According to Lakoff and Johnson, these oppositions are derived from experiential, bodily based gestalts.

But if they are right about the bodily basis of metaphor, the main contribution of Lakoff and Johnson's insights is to note that we need not accept as inevitable the current form and structtlre assumed by the comptlter. Computer systems are thought to work so well because of the generality of their basic grounding in binary logic. Computer programs, such as drawing systems or CCAD systems can be constructed because Cartesian notionjs of geometry map so well to this basic logic structure. In a computerized drawing system, lines are commonly defined as connections between points located anywllere in space, planes are bounded by lines, and volumes are spaces l~oullded by planes. The metaphors of containment, force, and balance outlined by Lakoff andJohnson are clearly very basic and map readily onto reductive Cartesian notions of space and geometry. But what if we reverse the priority of constructing drawing systems on

269 Metaphors and Machines

binary logic and instead begin with the concept of drawing? What sort of machine or computer system would support a non-Cartesian view of drawing? Other interpretations of the grounding of metaphor in bodily experience do not reduce geometry to conuainment, force, and balance but focus on the line as a boundary arc between near and far drawn by the sweep of the arm, on geometry as a generative device that has a life of ius own, and on the production of drawings as a journey by which we project an image into a situation and then return to find ourselves and the image changed.45 Such an account of drawing is more amenable to exploring the experience of manual drawing than reductive geometry is. We may then contemplate how a machine could be invented to facilitate drawing understood as such. By beginning with the bodily activity of drawing rather than the general mechanism of binary logic, a different kind of computer may be produced.

Such a line of inquiry is not so far-fetched. Some of the earliest computational devices controlled the deflection of an electron beam aimed at the center of a cathode-ray tube_early radar devices and oscilloscopes. Much research and development effort has been directed to suppressing the arc in favor of the straight line, and suppressing the natural propensity for mechanical and electronic componenus to exhibit states that are continuously variable and nonlinear in favor of discrete and linear states. The point is that once we accept the grounding of the computer in bodily metaphors, we are in a position to explore new aspecus of those metaphors, or new interpretations of the metaphors. Where this may lead in terms of computer systems design is unknown at this suage. Needless to say, there are many experiments exploring different paradigms of computation.46

No one would deny that the development of the computer has followed the best path technically, but what constitutes the best path is substantially a product of what fius within a complex matrix of other systems and inventions and what accords with our experience and expectations. According to the l.akoffandJohnson model, the cards are stacked heavily in favor of binary logic, and this is not because of some Platonic primacy of two-valued logic but because of our bodily engagement in a world of living, surviving, consumillg,

alla moving: "the massive complex of our culture, language, history, and bodily mechanisms that blend to make our world what it is "~7

According to the Heideggerian view and that of Derrida, critical theorists, and poststructtlralisLs, the impetus for developing modem technologies, such as the computer, and the reason they OCCtlpy sucll a place of privilege is metaphysical thinking. Can metaphysical thillking be explained by l.akoff and Johnson's account of bodily hased metaphors? From a Heideggerian viewpoint, the computer is caught tlp in a series of developments dating back to Plato and Aristotle. Technology's "essence" is our will to control, manipulate, manufacture, dissect, reduce, and "enframe." The development of the computer is abetted by this ethos. According to Hegel, Heidegger, and others, this takeover by Aristotelian logic_the reduction of reason to a manipulatioll, a technology_is to be contrasted with pre- Socratic thotlgllt, or dialectic, which celebrates opposition and colltraclictioll. The idea of oppositions seems to feature prominently in both pre-Socratic and Aristotelian discourse on reason. Both preSoctatic thotlght and Aristotelian logic are clearly "based in the hoc1y" in this sense (in relying Oll right-left and up-down oppositiOIIS). The major difference is that Aristotelian logic (at least according to the dominant tradition) denies its grounding in the world.~# According to the pre-Socratic view, reason is contingent and is based in contradiction_a dialectic hetween "is" and "is not." According to the Aristotelian view, reason is logic that can be di_orced from the day-to-day and manipulated as symbols on paper, and eventually on a computer. For pre-Socratic thought, the grouncl of reason is shifting and indeterminate. For Aristotelian thought, certainties underlie reason; there is structure that can be uncoverecl and decided. Metaphysics is that branch of philosophy that seeks Otlt the grounds for reason, the overarching principles, the framework that binds together all human knowledge. Needless to say, metaphysics is regarded with profound suspicion by most philosophical traditions since Nietzsche, though metaphysical thinking runs

271 Metaphors and Machines

deep in our institutions and in the forces that drive the development of technology.

There are four ways in which Lakoff andJohnson's view of metaphor impinges on the issue of metaphysics: first, the view that metaphor imbues all language presents a challenge to metaphysics; second, metaphysics can be described metaphorically; third, a challenge to the grounding of any understanding of cognition in the body must itself answer to the charge that it is metaphysical; and fourth, Heidegger charges that the entire metaphorical enterprise is metaphysical. I will consider the first three now, deferring the fourth until a later section.

First, Lakoff and Johnson address the issue of metaphysics indirectly. They devote considerable attention to challenging "objectivist" views of reason, which appear to deny the contingency of reason and insist on the existence of a literal, objective language independent of metaphor use. Having argued against objectivism, can Lakoff andJohnsoll be cast as relativists (subjectivists)? The answer is "no." I.akoff and lohl1sol1 also point to the contillgent and metaphorical nature of notions of subjectivity. Both objectivity and subjectivity arc metaphors within the Cartesian master metaphor of subject verstts object. In adopting the primacy of contingent metaphor use, Lakoff ancl lohnson's view attempts a direct challenge to metaphysics.

Second, their view provides a valtlable explanatioll of the mc taphvsical orientation_the quest for fatlndations. The strange alls in this argument is Derrida,49 wllo says the key to understanding metaphvsics rests in our propensity to see things in terms of plescncc and supplement. Metaphysics is not an appeal to deeply embedclc cl structure or pure transcendent reason so much as the simple recognition of things that are near and thir1gs that are far. For example, to say that there is both literal langtlage and metaphorical langtlagc is metaphysical. It implies that the literal is the true, essential, ar1cl immediate meaning; the metaphorical is the supplemental, aclclitional, unl1ecessary, and distant. The distinction between speech allcl writing is also metaphysical_Derrida's primaly example. Speech is near ancl essential; writing is a copy of speech. It is inciclental ancl distant. (Indirectly, Derrida is deflating the claims of metaphNsics

to transcendence by showing its grounding in the everyday issue of proximity. )

Applying Lakoff andJohnson's (bodily) point of view to Derrida's argument, metaphysics is grounded in the most basic metaphorical image schema we can imagine, that of inside versus outside_the containment schema. Derrida's identification of presence versus supplement is little more than an identification of the basic schema by which we identify what is contained and what is outside, what is within our reach and what is outside it. As already discussed,Johnson describes this schema at great length and shows both its grounding in bodily experience (ingesting food, discovering limits to the reach of our arms, and so on), and its prominence in language. We can extend the containment schema to a consideration of the preSocratic, Pythagorean identification of binary oppositions. The oppositions between straight and curved, up and down, and right and left are not symmetrical. As Derrida has shown us, there is always a privileging. In other words, as Pythagoras said of these oppositions, they pertain to the limited and the unlimited. There is only one way for a line to be straight. A straight line is present. A curved line is a deviation from straight, and there are unlimited ways for a line to be curved. A curved line is supplemental to the "ideal" of a straight line. So the containment schema is based in bodily experience, it is ubiquitous, and it is implicated in pre-Socratic binary dialectic, in ordinary and abstract language (according to Lakoff andJohnson), and in metaphysics (according to Derrida). In view of this, our propensity for metaphysical thinking should come as no surprise. We think metaphysically because we are grounded, bodily beings, and our bodily experiences lead us to see things in terms of containment.

Of course, Derrida's project is to show the tenuous nature of metaphysical thinking, how the pri\ileged term in a binary opposition can be shown to be the lesser term, how speech depends on writing (protowriting), and how literaHangtlage depends on metaphor. Derrida suggests a return to the pre- Socratic view of thought (lo,~;os) as being in constant flux. He does this by showing us how committed we are to metaphysical notions, while at the same time he removes their certainty.

l

i

l

We have shown how one aspect of Derrida's account of presence and supplement can be explained in bodily terms. This brings us tc the third meeting point between the metaphorical and the metaphysical. What does Derrida have to say about the bodily basis of thought?

Derrida already posits a sophisticated deconstruction of the containment schema by showing that what is present depends on what is supplemental, not as a universal rule, but through many specific cases pertaining to language use_meaning, naming, metaphor, writing_and philosophy, specifically Heidegger's notions of Being. Given this, a Derridean critique of Lakoff andJohnson's view of metaphor would not be to refute it, but to play with it, to show it as contingent and self-referential.

To date Derrida has not specifically addressed the writings of Lakoff and Johnson, but a critique comes in the form of Derrida's identification of the present and the supplemental in Heidegger's writing about the hand, which is a bodily metaphor of some potency in Heidegger's writing.50 In Being and Time and elsewhere, Heidegger makes great use of the hand as a means of distinguishing between the immediacy of our involvement in the world and our abstract theorizing. There is the readiness-to-hand of immediate involvement and the presence-at-hand of the world of measurable objects. Derrida thinks Heidegger is unwittingly succumbing to metaphysical thinking in uncritically according to the hand this immediacy and power. If the charge of metaphysical thinking can be leveled at Heidegger over the issue of the body, then the same charge certainly applies to Lakoff and Johnson. They privilege the body without apparently acknowledging the contradictions and reversals this entails. To trace language and metaphor use back to the immediate realm of bodily experience is a metaphysical exercise. It assumes a base, a solid foundation for understanding reason. A~s Derrida points out, we can only ever present our arguments in such a way, but postmetaphysical writing presents such ideas as "under erasure." We indicate the contradictions they entail and show their provisional status. Of course, 1,akoff and Johnson are writing in a scientific and analytical context rather than a philosophical or literary one. Although their theorizing is far from objectivist, it is prag

matic. The rules of the game are different from those enjoyed by literary deconstruction. Their thesis about the bodily basis of metaphor opens up avenues for exploration, and the ultimate arbitrator of the validity of these ideas is their usefulness.5'

Other Derridean arguments can, however, be advanced against Lakoff and Johnson's view. Johnson identifies containment, force, and balance as bodily based image schemata. He shows how they are each implicated in both everyday argumentation and in Aristotelian logic. In keeping with Derrida's deconstruction of concepus of speech and writing, we can show that the notions of containment, force, and balance are not simply borrowed from the physical world of human action and applied to formal logic but depend on notions in formal logic. How is this so? Containment, force, and balance were obviously not invented by logicians to be applied only to the world of physical action later on. Rather the terms we normally take to apply to putting something inside a container (containment), pushing against a physical object (force), and ensuring that we do not fall over (balance) are but manifestations of more-basic notions of containment, force, and balance. These are preembodied notions. As preembodied, these notions may be partially lost to everyday language but are partially uncovered in the consideration of the abstract world of Aristotelian logic.52 That these image schemata are present in logic, which makes no overt claims to being embodied, provides evidence at least that there are other nonembodied uses of the terms "containment," "force," and "balance."

The best proponent of this kind of deconstructioll is Heidegger, particularly in his notion of the primordial "in."53 For Heidegger, the spatial "in" of containment is subservient to a primordial notion of "in" as involvement. There is the nonspatial "in" of being-in-tlleworld, being in a good mood, being in love. Seen in this light, 1,akotl and Johnson's notion of containment is subservient to the more primol-dial notion of involvement. Prior to our bodily experience of containment is our being-in-the-world, an altogether more primal~ and important concept. Similarly, Heidegger offers a revision of notions of causality, whicll for l.akofl and |olluso'~ is rcl.ltetl to [lit bodily experience of force. For Heidegger' causality is subsel~iellt to care. From our being-in-the-world, we direct Otll attelltiOII Withit

l

l

1,

l

a region of concern. Notions that we may cause something to happen and that we may exercise control over a situation are derivative of this more-basic understanding of our place as exhibiting care. As a further illustration of the subservience of notions of force, Heidegger points to our ability to already be where we want to go. Before I move across the room, I am already there ontologically by virtue of my involvement in a sphere of concern that includes my "destination." It is only because I am already there that I am able physically to go there.

These argumenus are obviously counter to those proffered by LakoffandJohnson. Heidegger argues that there is a more basic experience than embodiment. Even orientation has a grounding prior to our bodily experience of orientation. So Heidegger conspires with his later opponent Derrida in identifying a preembodied mode of being. Whereas Heidegger's identification of preembodied experience could be construed as yet another instance of discovering a foundation (not in the body, as for Lakoff and Johnson, but experience prior to the body), the preembodied has the appearance at every turn of being undecided. It is a fluxional involvement that defies pinning down. Heidegger's primordial concepts are not foundations but excursions into pre- Socratic concepts of contradiction, flux, and play. How else could we characterize being-in-the-world?

In a strange way, this Heideggerian-Derridean critique of embodiment also finds some support from Davidson. The use of a term such as "in" is never limited to the metaphor of containment. A floor can be part of a building or apply to the earth or any solid grounding, a file is not just a cardboard folder but storage for information, and a desktop is not just a place to rest books when I am reading them but a computer display. In the same way "in" is not simply the spatial relationship between an object and a container but applies to whatever uses our language community is able to find for the word. In this sense, all words have a "prior" meaning that goes beyond any notion of an "original" meaning. The small word "in" is a placeholder in a complex of varying contexts and practices, best captured h)' tiltmlOtiOn ol USC. Tl\iS iS consistellt with pragmatic theories of mt aning. As a variation on this pragmatic theme, a Derrideall reading of the word "in" places it, along with all other words, in a C0111

plex play that is difference and deferral. In either event (whether we adopt the Davidsonian or the Derridean understanding), the meaning of "in" does not reside in the association we establish between an image schema of containment and some new domain of abstract reason. At every turn, Lakoff and Johnson's claim for the primacy of bodily based metaphors is under challenge. Derrida, Heidegger, Lakoff, and Johnson each challenge metaphysics, though LakoffandJohnson's insistence on the primacy of the body appears to undermine their own intentions here.

But then again, Lakoff andJohnson's notions of the embodiment of metaphor have uses_provided we see embodiment itself as a metaphor and not as having some absolute foundational status. The uses of Lakoff and Johnson's notions of embodiment include restoring dignity to the body in philosophical discourse, refuting certain Cartesian dualisms, rehabilitating reason as grounded in worldly concerns (as opposed to transcendence), and showing how we can see ourselves in our technologies.

Metaphor challenges metaphysics, but what about the scientific basis of computer systems? It is commonly thought that designers of computer systems may use metaphors as heuristics, as ways of talking about systems, interfaces, and so on, but the computer is best understood scientifically.54 If we strip away the devices designers use for talking about design, such as metaphor, we will arrive at the scientific underpinnings of computer systems. But even if we hold to this view, we find that these "scientific underpinnings" are also metaphoric. This strikes another blow to the idea of metaphysical underpinnings beyond metaphor.

One of the foremost proponents of the view that science is driven by metaphor is Hesse.55 If she is right, then whichever way we look at it the computer is pervaded through and through by metaphor. As I discussed in chapter 6, the link between science and metaphor is forged through the idea of a model. We will now develop the idea of models further. According to Ricocur, "The function of a model is to describe an unknown thing or a lesser-known thing in terms of

277 Metaphors and Machines

a better-known thing thanks to a similarity of structure.'t56 In describing scientific activity, it is common to distinguish between the thing observed and the means at our disposal for explaining it. The thing observed is described in an observation language_a language that scientists develop for describing what they observe_for example, "light and dark bands appeared when light is passed through two parallel slits onto a white card." Then there is the model or theory for explaining the phenomenon_the wave model of light. Hesse actually describes these in terms of a primary and a secondary system. The primary system is the observed phenomenon and the general terms used to describe it_the behavior of light; the secondary system is the system of explanation_the wave model. According to Hesse, "in scientific contexts the primary and secondary systems may both be highly organized by networks of natural laws."57

In the case of computer systems, computation provides many models for other phenomena. There are computer models of natural processes, chemical reactions, weather, the behavior of populations, and so on. But there are also models of computation, and these provide its scientific base. As we have seen, the Turing machine is one such model.58 A simple machine consisting of paper tape, a mechanism that can move the tape backward and forward, and a device for reading and writing to the tape make up a universal machine for simulating all algorithmic processes, including those of any computer. The Turing machine model incorporates theories about properties of algorithms, such as decidability, consistency, regularity, and determinism. The theory is generally known as automata theory.59 In this case, the primary system is the behavior of an algorithmic process, the behavior of a computer; the secondary system is the system of explanation, the Turing machine as a model.

How are the primary and secondary systems linked? Hesse contrasts the deductive with the metaphoric view of science. According to the deductive view, the thing explained (say, the interference pattern of light shone through two slits of paper or the observation that an algorithm is running in a loop without generating an outcome) is deducible from the scientific explanation (the wave model of light or the theories of the Turing machine). That is, science posits explanatory models, and given the model we should be able to derive

the phenomenon observed. The explanatory model must be capable of prediction. In order to do so, the explanation must contain at least one general law that must not be empirically falsified to date. In the case of the wave model (and other models), one such law is that light intensity diminishes inversely to the square of the distance from the source.

As Hesse indicates, there are problems with this deductive view. Since Popper, it is widely accepted that obsenation is theory-laden. The observation language already betrays a commitment to some theory or other. The observation that an interference pattern was observed presumes a theory of interference_patterns reinforcing and canceling each other out in some way. The observation about a runaway algorithmic process presumes some notion of looping. Further, the law or theory for predicting an observation from a model is problematic. The fit between the model and the observatiO'I is always approximate. How approximate is a matter of interpretatioll, and there appear to be no laws for determining what constittltes a close fit between the two systems. The deductive view of models offers no explanation of the link between the primary systelll ancl the secondary system. This is the case for computation as mtlcll as for models of light. The difficulty of making connections between Turing machines (and automata theory) and the behavior of computer programs is legendary. Very few programmers are willing or able to make use of it in any practical sense to predict the behavior of their programs.

By way of contrast, what is the metaphorical view of scientific models? According to Hesse, a scientific model is a metaphor. The primary system, the observation langtlage, is the first term of a metaphol~. The model is the second teen. The wave model of light can be cast in metaphoric terms: light as waves. The Turing machille model presents the metaphor of the computer as a machine for readhlg and writing to tape. Scientific models can therefore be accotlllted for according to the interaction theory of Black and others. But there are important differences between metaphors in science and metaphors in literature. At first blusll, metaphors in science appear to support the similarity view of metaphor. Scientific metaphol~s appear as ~imiles. anct thes- C;lil he recitlcecl to literal statelllellts. I hel e are t~vo ways that this appears to be the case.

First, according to Black's interaction theory of metaphor, the metaphor creates the similarity rather than bringing to light some preexisting similarity. In contrast, in science it is generally assumed that the model reveals some preexisting similarity. Light simply does behave as waves, independently of any models or metaphors we care to fabricate. Computer programs simply do behave as Turing machines. Second, whereas in literature any metaphor offers something illumillating_Juliet as the sun, a rainbow, an auttlmn mist, a thtinderstorm_in science not every metaphor is applicable in every sittlation: "[N]o model even gets off the groulld unless some antecedent similarity or analogy is discerned between it and the explanandum [the thing explained]."fi_

But Hesse indicates that in spite of these differences the substitution theory does not hold even for metaphors in science. In projecting a model/metaphor onto an observational situation, the scientist never knows how far the comparison extends. There is never a definitive list of similarities that exhausts the comparisc>n, ancd it is the unexpected extensions of the compal-isons that prolllott the development or rejection of a model: "[A]s long as the model is under active consideration as an ingredient in an explanation, we do not know how far the comparison extends_it is precisely in its extension that the fruitfulness of the model may lie."fi' Accordblg to Hesse, there are two further differences between metaphors in science and those in literature anc1 poetry. First, metaphors in literature and poetry are initially striking and unexpected. They are nc~t meant to be analyzed in pedantic detail, and such analyses destroy the metaphor. In science, models/metaphors may be only initially ullexpected. They are not meant to be peculiar but are meallt to be cc~llsistent and tightly knit intenlally. For a model/metapllor h1 science, contradiction is not desirable. Science seeks to finct the "perfect metaphor."

Secolld, in literary language, a metaphor is not put forwal~ci as explaining something: "[I]n literary metaphor in general there is no hint that what is metaphorically described is also thereby explained."fi2'But in science, moclels are cotl__tecl as explanaton. What is explamatioll.~ The deductive \iew of science assullles that the explanatol y power of a model resides in correspondence r ules

joining the two languages: observation and explanation. As has been pointed out by many commentators, this is a problem for the deductive view. On the other hand, according to the metaphorical view, there are no correspondence rules. Furthermore, there are not two languages, the observation language and the explanation language (the language of the model). There is only one language, the observation language, which is continually being extended by metaphoric uses. For example, it is extended into the terminology of the explanation_light is described not only in terms of light and dark bands but in terms of waves, frequency, amplitude, and interference (from the wave model/metaphor). As a further example, under the influence of the Turing machine model, computer programs are described in terms of symbols, start conditions, and rewrite rules (the action of the read/write

device in the Turing machine). So the use of models/metaphors in science is no different from the use of metaphors in literature or in day-to-day language. According to Hesse, this process of extending language through metaphor use is not necessarily well understood, but it is a process of language generally and is not peculiar to science.

According to the metaphoric view, predictions become possible by the shifts in meaning and extensions of vocabulary brought about by metaphor. These predictions may or may not turn out to be empirically verifiable: "They will however be rational, because rationality consists just in the continuous adaptation of our language to our continuously expanding world, and metaphor is one of the chief means by which this is accomplished."fi3 So, following Hesse, even if we regard the computer as a product of scientific reasoning this does not exempt it from its grounding in metaphorical thinking.

We have appealed to the usefulness of metaphor. What is the efficacy of metaphor use in understanding information technology? We can summarize the implications so far. There are several means by which the link between metaphors and technology can be forged. Technology provides metaphors through which we understand phenomena,

metaphor is implicated in the setting of problems, dependencies between technologies can be understood as metaphorical, and metaphors provide a basis for evaluating technologies.

Technology provides metaphors through which we understand phenomena_the mind as computer and the house as a machine_ and technologies in turn are understood through other phenomena. Basalla provides an example of the metaphorical relationship between technology and biology.fi4 He notes that early philosophers, such as Aristotle, wrote extensively on biology but did not make much use of mechanical metaphors. With the Renaissance development of machines, Renaissance writers drew parallels between the mechanical and the biological: "Structures and processes in living organisms were described and explained in mechanical terms."fi5 In the nineteenth century, the relationship was reversed. With developments in geology and Darwinian evolution, technology was described organically. Basalla cites the example of various tables and charts drawn up during this time that placed artifacts (such as axe heads and weapons of tribal cultures) into evohltionary trees, with genera, species, and varieties. This is still common practice toclity. Technologies are described biologically, and biology is somethlles understood in terms of technology. We see similar reciprocity between technology and organism in the case of computers and mind. As pointed out by Ryle, Schon, and Sternberg, the technology ot mechanization has also furnished us with the metaphor of the mind as a machine, but certain machines (computers) are also uncterstood in terms of mind.fi6

Technology as a source of metaphor can also operate in more subtle ways. Metaphor provides a way of accounting for the power accorded to technology by such writers as McLuhan and Ong, who, as we saw in chapter 1, assert that the technologies of literacy and print have contributed substantially to the primacy of method, individualism, and notions of objectivity, leading the way for Descartes, the Enlightenment, and modern science.fi7 According to Ong, people from literate cultures (like ours) have so "interiorized" the technology of writing "that they organize, to varying degrees, even their oral expression in thought patterns and verbal patterns that they would not know of unless they could write."fi8 Our deference to the

syllogism illustrates this. Ong makes an even stronger point. The technology of print has persuaded us to think of mind in spatial and "objective" terms.

By removing words from the world of sound where they had first had their origin in active human interchange and relegating them definitively to visual surface, and by otherwise exploiting visual space for the management of knowledge, print encouraged human beings to think of their own interior conscious and unconscious resources as more and more thing-like, impersonal and religiously neutral. Print encouraged the mind to sense that its possessions were held in some kind of inert mental space.69

By this metaphorical construction, McLuhan and Ong are positing grand master metaphors that influence the way we see the world. McLuhan also ascribes such movements to the engagement of the different senses, particularly hearing and vision. So print also gains its metaphorical power due to its involvement in the sense of sight.

The second link between metaphors and technology is that metaphors appear to generate problems that technologies are presented to solve. Problems emerge as the entailments of metaphors. The view that technology and metaphors interact puts to one side any idea that technologies arise as a response to some abstract need or problem. It suggests that technologies, metaphors, and problems are interrelated. This relationship is made clear by Schon,'_ who says that different metaphors promote different problems. Technologies are commonly advanced as the solutions to particular problems, yet it is apparent that the technologies have brought about the metaphor shifts that define and even create the problems in the first place. Thought is shaped by the technologies in such a way that the technologies are readily accommodated. Information processing provides a good example of this complex play between technology, metaphor, and problem.''

When we operate with the information-processing metaphor, certain problems present themselves. For example, problem solving and design, seen as information processing, lead to a concern with accuracy of information transfer_minimal disruption to the flow of information from the "coder" to the "decoder." In architectural and engineering design practice, this may be seen as the problem

283 Metaphors and Machines

of accurately communicating information from the designer to the manufacturer or tradesperson. The technological solution might be better CAI) (computer- aided design) database techniques or morecomplete specifications. However, in an age in which the metaphor of information processing had less force, we would expect other problems to have presented themselves. So problems (such as accuracy) emerge from the metaphors provided by the technologies (information processing).'2

We can also see that changing the dominant metaphors redefines the problem. There are other metaphors of human communication, cognition, and design. For example, design can be seen as a process of enabling within a community of expertise. The problem of accurate information transfer then becomes a problem of facilitation. How can clients, consultants, authorities, contractors, and tradespeople be enabled to do their job or carry out their roles better? This new way of looking at the domain may result in a spectrum of actions, ranging from the adoption of new media or new ways of using computers to different modes of education within the professions and trades. The solution may not involve technology at all but rather new practices.

The third linkage between metaphors and technologies is that dependencies between technologies can be understood as metaphoric. Metaphor features in an understanding of important dependencies among technologies. In the history and "evolution" of a family of artifacts, such as pottery jugs or certain building types, certain features persist that seem to serve no functional purpose. "Skonomorphism" is a term coined by archaeologists to designate a "nonfunctional" feature of a design that derives from some precul

sor to the current design.73 For example, the projecting ends of wooden beams appearing originally in timber buildings reappeared as ornamental dentils beneath the eaves of Greek temples built of stone. Early plastic laundry baskets were fashioned in imitation of the reed and wood construction of the earlier artifacts. The phenomenon relates mainly to changes in materials, but skouamorphism can be evident in any aspect of an artifact or a technology. We under

stand and develop new technologies and new designs in the ligllt of our experience of existing technologies and designs. This is a

1

metaphoric phenomenon. The early technology sheds light on the new technology. New technologies, inventions, and designs are seen in terms of existing technologies, inventions, and designs: stone construction is seen as carpentry; plastic is seen as reeds and wood. The phenomenon is certainly evident in the design of computer systems. As I stated at the beginning of this chapter, the success of the personal-computer interface is largely attributable to the deliberate use of particular organizational metaphors (desktop, folders, windows, and trash can). Certain computer-aided drafting systems make extensive use of metaphors wrought through a consideration of manual drafting tools: the screen cursor looks like a pencil or a paintbrush.74

When we consider the interactive nature of the terms of a metaphor, we see that the relationships among technologies are not static. In the same way that "house is a machine" can inform us about both terms ("house" and "machine"), new technologies shed light on old methods of working. Construction in stone tells us something new about construction and structure. Understanding stone's resistance to compression and its susceptibility to shear informs us about the unique behaviors of timber. We look at timber construction in a new light. The plastic laundry basket informs us about the unique properties of reeds and wood. We understand a computerized drawing system through our knowledge of manual drawing, but a computerized drawing system also informs tlS about drawing in general. There are new operations that we formerly did not consider part of drawing (stretching, inverting, smoothing). We look afresh at conventional drawing tools, and we learn new things about creating images. We may value the uniqueness of manual drawing in a new way_seeing drawings not merely as points and lines but as extensions of bodily movements, something CAD drawings clearly are not. We may presume that this dynamic relationship between technologies informs the development of new technologies, inventions, and designs.

The fourth link between metaphor and technology is that metaphor provides a basis for evaluating technologies. Evaluation proceeds on the basis of the efficacy of the metaphors that the technology brings to light. How does this occur? According to this

285 Metaphors and Machines

model, evaluating a technology involves identifying the dominant metaphors that influence the user at the work task and that are attributable to the presence of the technology. This involves understanding the experiences of the computer user. The next step is to identify the sets of problems these metaphors define. We can then evaluate the efficacy of these problems in terms of the extent to which they facilitate the work task, the extent to which they represent the imposition of "false" problems that disappear once new metaphors are adopted, the extent to which problems are a product of the metaphors imposed by the technology, and the extent to which problems change their definition once new metaphors are introduced.

The actions that follow from this analysis may include introducing technologies, practices, and interactions that provide more-enabling metaphors; arguing for the adoption of new metaphors for understanding tasks (new ways of looking at tasks); recommending new metaphors through which the technology can be understood; and redesigning, changing, or abandoning the technology in particular situations.

The critique of information processing given above follows this kind of analysis. On a smaller scale, one could apply this method of critique to some aspect of a computer system, such as the screen cursor in a computerized drawing system. A system design in which the cursor looks like a pencil presents problems of control. Screen pencils are notoriously difficult to use. We may expect the cursor t0 work like a pencil only to find that we are unable to use broad strokes or change the thickness of the line as we draw. The problem maY disappear once we see the tool as something else, perhaps as a fine paintbrush or an etching tool. We may also circumvent the problem of controlling the cursor in a pencil-like way by analyzing the metaphor of control itself. The idea of control is heavily promoted through the technology of computation. What appears on a computer screen is so heavily mediated by electronics that we do not have immediate access to the work surface (the screen) other than through the control of keyboard, mouse, and numbers. When we use a physical pencil, we do not so much control the pencil as make use of a tacit understanding, or familiarity, with the tool within a

context of skilled practices. The problem of the pencil-like screen cursor can be recast as a problem of an inappropriate expectation due to an inevitable distancing by the technology.

The metaphor of distancing may suggest that the computer medium is useful for performing abstract operations, such as geometrical transformations. What becomes important is not getting the marks onto the screen but how one manipulates them once they are there. This may involve abandoning the pencil cursor or changing the screen display in some way that makes the mediated nature of the pencil cursor apparent. This is just a simple example of how an understanding of the ability of technologies to furnish us with metaphors can be brought to service in evaluating aspects of technologies in use.

In summary, metaphor and technology converge in several ways. Technology functions as a source of metaphors, and technologies are understood metaphorically through other phenomena. Metaphors, problems, and technologies are interrelated. Metaphors set problems that technologies are commonly put forward to address. These technologies in turn promote metaphors that set the problems. Technologies also provide metaphors of each other. This particular technology or artifact is seen as that previous technology or artifact, and the reverse also applies. We revisit previous technologies in the light of what is now available. Finally, metaphor provides a mealls of evaluating technologies.

Metaphor proves its efficacy in these areas, but what is its role in clc sign? Before we can address this question, we need to deal with I leiclegger's criticism of metaphor, which brings to light further issues that impinge on design.

\N'c have explored at length Heidegger's view that technology brings ahotlt a covering over of Being in favor of a focusing on beings. In other words, thanks to technological thinking dating back to Plato, we have lost sight of what it means to be. We now focus on simple entities, objects as causally related, and human thought as a matter of manipulating propositions according to causal laws. In the pre

287 Metaphors and Machines

Socratic mode of thinking, to which Heidegger wishes us to return,

we let things be in their essence.75 That is to say, we allow things to be different in uniqtle ways, to fit their context, to be their OW11 causes. This is not to exclude causal thinking as a valid mode of investigation, as long as we recall the derivative and provisional nature of such thinking. In his later work, Heidegger indicates how the technological enframing to which we are now committed is inescapable and not "caused" by humankind so much as "sent" by Being. Being covers over and reveals itself at different times and in different ways throughout history. Ns we have seen, Heidegger's later thinking is controversial and complex and borrows substantially from the German theological and mystical traditions. Heidegger's life's work therefore focuses on the matter of Being. How do Heidegger's investigations into Being relate to metaphor?

Heidegger begins his first major work, Being and Time, with an argument for the resurrection of interest in the verb "to be"_the "sum" ("I am") of Descartes's foundational proposition "Cogito, ergo sum" ("I think, therefore I am"). As discussed earlier, metaphor in language is the linking of two terms through the copula "is"_the computer is an intelligent assistant, and a house is a machine. It would seem that metaphor, then, is central to Heidegger's project to cdevelop an understanding of Being. Unfortunately for the cause of metaphor theory, Heidegger attacks any claim to the primacy of metaphor.

Althougll Heidegger does not mention metaphor in Bein,< and Time, his later attack on metaphor is consistent with his explanation of the "as" stmcture in Being and 7ime. The preposition "as" is the other connecting term used in metaphor_the computer as an intelligent assistant ancd a house as a machine. In Being and 7ime, Heidegger distinguishes between the hermeneutical and the apophantic "as." The hermeneutical "as" is the immediate recognition or projection that derives from being-in-the-world, encountering the ready-to-hand as part of our circumspective awareness, our dealings in which we exhibit concern.76 We have this understallding in advance. This is the primordial, premetaphorical use of "as." On the other hand, the apophantic "as" is the ontic, representational way in which we assert something as something.77

Heidegger's later assertion of the primacy of a prelinguistic mode of understanding that eschews metaphor is consistent with this. Dreyfus summarizes Heidegger point: " [A]t an early stage of language the distinction metaphorical/literal has not yet emerged."'8 This is not to deny that there is metaphor. Heidegger is simply asserting that there is a more-basic understanding that precedes metaphor use. There is a more-basic use of "as" that is not metaphoric.

Heidegger's argument can be taken two ways. If we take metaphor to be a mere phenomenon of language, then we have to agree that there is prelinguistic, premetaphorical understanding. However, if we take Lakoff and Johnson's view about metaphor as understood in terms of experiential gestalts, then metaphor runs deep in human experience_"as our mode of being-in-the-world or our way of having-a-world," according toJohnson.79 According to this view, Heidegger's commentary on the "as" structure points to a redefinition of metaphor to include ontological as well as ontic use. There is hermeneutical as well as apophantic metaphor. By this interpretation, metaphor runs deep in our involvement in the world, and it is a prelinguistic phenomenon, as Lakoff and Johnson suggest.

Elsewhere in Being and Time, Heidegger describes this hermeneutical, prelinguistic "metaphor use" in his description of projection. This description covers both the way we deal with everyday situations and with analytical study. According to Heidegger, all interpretation begins with a background that we take for granted (a fore-having). There is also always a point of view (a fore-sight), and there are expectations (a fore-conception). This is Heidegger's "three-fold structure" of interpretation_fore-having, fore-sight, and foreconception.80 This structure can be taken as a description of metaphorical projection of the kind Lakoff and Johnson describe. To interpret a situation is to engage in metaphorical projection. It is to see the current situation as another situation. Every experienced moment is a candidate "second term" of a metaphor for the present moment. It is to project a point of view and an expectation into a situation. There is therefore ample support within Being and Time for the primacy Lakoff and Johnson and other metaphor theorists accord to metaphor_as long as we see metaphor as more than a phenomenon of language.

What of Heidegger's specific criticism of metaphor? In his later writing, he asserts that to interpret certain statements as metaphoric ~s to show a commitment to metaphysics: "The metaphorical exists only within metaphysics."8' Wood makes Heidegger's argumellt clear.82 According to Heidegger, to explain his own work as comprised of metaphors is to maintain the possibility of the literal. To assert as much is to maintain the presence of some absolute. It is to get back to old metaphysical notions of language as reference_a stable world to which words refer. Heidegger's attack comes in the context of his discussion of seeing and hearing as ways of thinking presented in Der Satz vom Grund ( The Principle of Reason).83 Lest the reader think Heidegger is merely being illustrative in characterizing seeing and hearing as ways of thinking, Heidegger asserts that he is not speaking metaphorically. Nor is he speaking literally. To assert as much is to maintain that there are sensory levels of experience (seeing and hearing) as distinct from intelligible (thinking) ones. According to Heidegger, such a dichotomy rests on Cartesian and metaphysical distinctions between body and mind. According to Wood, to assert that Heidegger's reference to thinking as seeing and hearing is metaphoric "would rest on the assignment of seeing and hearing to the sensory level and thinking to the intelligible or nonsensory."84 The observation about perception given earlier_that we do not hear noises that we then translate into meaningful sounds, but that we directly hear cars, birds, and bells_takes a more radical turn. To hear and see is to interpret, which is to say, to think. In breaking down the mind-body dualism, this argument, also developed by Wittgenstein in Philosophical Investigations,85 is an attack against metaphysics itself, and by implication against metaphor. But, according to Wood, it need not imply a serious challenge to metaphor.85 The argument relies on drawing a general conclusion fron~ particular examples. To assign Heidegger's discussion of seeing ancl hearing as thinking to the realm of metaphor may be an "error" in this particular case, because he is making a point about breaking down the sense-thinking dichotomy, but not every categorization of an utterance as metaphoric is in "error." The criticism against metaphor need not always hold. To Wood's argument in defence of metaphor, we can add that Heidegger is again assuming metaphor to be

a phenomenon only of language, and a superficial phenomenon at that_ideas strongly refuted by the contemporary metaphor theorists referred to so far.

The second occasion on which Heidegger eschews metaphor is in his essay Letter on Humanism. Metaphysics is that tendency within philosophy and reflection to mask Being in favor of beings, in other words, to miss the basic questions of Being that the pre-Socratics pursued and to see the existence of beings as important_the dominant aspect of the philosophical tradition since Plato. How is metaphor metaphysical in this sense? As an example, Heidegger at one point describes language as "the house of Being." On first reading, this comes across as a metaphor_a rich, evocative, enigmatic metaphor. However, in his Letter on Humanism,87 Heidegger explicitly says that the expression is not to be taken metaphorically: "The talk about the house of Being is no transfer of the image of 'house' to Being. But one day, we will, by thinking the essence of Being in an appropriat(ing) way . . . more readily be able to think what 'house' and 'to dwell' are."8R One point of the essay is that we do not understand what it is to dwell. Our need and ability to dwell has been covered over by the philosophical tradition. For Heidegger to dwell is to "be-in-the-world." In recovering the primordial meaning of "dwelling" and "house," we are returning to an immediate ontological use of the word "house." Language as the house of being should not, therefore, conjure up an image of a house with something called "Being" in it, and the whole thing bearing the label "language." Nor, obviously, should the phrase conjure up some lists of features attributable to the three entities language, house, and Being so that we can compare them. If we have attended to Heidegger's carefill arguments in his essay, then we are already familiar with his newly rehabilitated notion of dwelling and house. The statement is indeed not to be taken metaphorically. The statement relies on a more-primordial notion of house that transcends the literal/ metaphoric.89

Derrida takes Heidegger's suspicion of metaphor as a key theme in his argument about metaphor. Derrida's argument follows four stages. First, every attempt within philosophy to explain mc taphor is itself imbued with metaphor, and every description of metaphor

depends on philosophical ideas.90 A mutual dependence is therefore in play_philosophy depends on metaphor, and metaphor depends on philosophy. Second, Derrida rejects meaning as depending on reference (that is, this word refers to that object) in favor of meaning as dependent on difference. According to Derrida, to distinguish between words and meanings is to affirm the tenets of metaphysics and the possibility of the literal. With the disappearance of the idea of the literal, metaphor also withdraws from the scene. But, according to Derrida, we use metaphysics against itself. We borrow from metaphysics. Deconstruction is that strategy by which we construct arguments, inevitably depending on metaphysical assumptions, but we write "under erasure," showing the provisional nature of our terms. It is never possible to get outside of metaphysics. One can only hope to work within it, but in a way that is subversive of the arguments of others and of one's own arguments. So too with metaphor. We use it, refer to it, and subvert it.

Third, metaphor itself is located within the play of differences. It is not based simply on the opposition of sensible/nonsensible (the senses and the intellect), as suggested by Heidegger, but on a whole range of distinctions. In the literature on metaphor, notably from Nietzsche, there is talk of wear and tear (worn-out metaphors), life and death (dead metaphors), and use and profit (coins that are no longer of accowlt).9~ Derrida is keen to show that metaphor does not depend on one opposition but on many. The very idea of metaphor is slippery. Whereas Heidegger looks for dwelling and a sense of stability, of being at home, Derrida indicates permanent dislocation.

Fourth, Heidegger talks of the withdrawal of Being. This is Being's doing. Yet Being has accomplished this in such a way (a series of guises, turns, modes) that we are tempted to describe Being througl1 all the ruses of metaphor. There is no nonmetaphorical sense in which we can talk of the withdrawal of Being. As in all Derrida's arguments, this brings us back to the primacy of difference and deferral.

In summary, lleidegger addresses the issue of being. At first reading, this seems to implicate metapllor, which, with its heavy investment in the copula ("is"), appears somehow central to Otlr

being-in-the-world, as Lakoff and Johnson suggest. But Heidegger eschews talk of metaphor as belittling his insights into primordiality and prelinguistic understanding. He seems though to have a very linguistic-centered view of metaphor. The view of metaphor of LakoffandJohnson and others accords more with Heidegger's notion of the hermeneutical "as"_the prelinguistic recognition of this as that. For Heidegger, the question of Being is resolved by our attending to our home, or dwelling in the world, with all the implications of stability. Again, for Heidegger, the reference to home is not a metaphor but strikes at the essence of Being. Derrida, on the other hand, thinks Heidegger is caught up in all the ruses and turns of metaphor in his accounts of Being. Most sign)ficantly, whereas Heidegger seeks stability, Derrida seeks permanent dislocation_the constant reminder of uncertainty. This latter insight points again to the workings of metaphor and the role of difference and deferral.

Before returning to the matter of difference, I will review what metaphor tells us about design. It is clear that metaphor provides insights into the workings of the design process. Design can be characterized as generating action within a "play" of metaphors.92 We see the design as particular things during its development. This is evident in the case of geometrical manipulation. We see the C011figuration of marks on the drawing board or computer screen as triangles, squares, circles, diamonds. The things we see suggest "problems" (or opportunities). We know about a square and the manipulations we can perform with it. These are the entailments of the metaphor. So perhaps we draw in a diagonal. New shapes and figures emerge. We cannot help ourselves. Our experience is constantly bringing metaphoric projections to bear on the current situation. It is constantly undergoing transformation in the light of the current, changing situation. Of course, even when playing with geometry, the process is never entirely geometrical. Shapes are never just shapes but also swlbursts, leaves, rays, clouds, tokens, wings, crystals, window panes, and symbols. And each metaphor suggests new action.

Presumably design communities have access to certain metaphors that are considered enabling within particular domains of action. So there are different ways of understanding a design in progress as the designer shifts from one metaphor to another. Different metaphors have currency within different design communities. For example, an architect may see spaces as having the properties of a fluid ("space as fluid" metaphor). This metaphor suggests and opens up possibilities (or problems) of "shaping" spaces, improving their "flow," and identifying their "source." If a designer is operating within a "form as mass" metaphor, it may lead the designer to consider "emphasis," "giving weight" to an element, and providing "balance." Some of the metaphors used by architects enable them to see their designs in terms of focal points, concepts, points of departure, themes, ordering principles, and types. There are also various analogs_shells, pine cones, praying hands_and precedents that arise as metaphors. Certain metaphors are unique to certain design disciplines, such as architecture_providing a sense of address, defining the street, interface, accentuating, articulating, providing a transition, providing an interplay of spaces. But metaphor play is also evident in the supposedly "practical aspects" of design_ a building as a system, machine, energy sink, circuit, ledger item, resource, or liability.

So too in the design of software and computer systems. We find overarching metaphors through which problems are set, such as the computer as intelligent assistant, theater, environment, ubiquitous resource, and so on. As long as we see the computer as an intelligent assistant, we set ourselves the agenda of codifying knowledge, replicating human thought, modeling cognition. To see the computer as theater is to think in terms of scripts, actors, spectacle, engagcment, and so on. These metaphors may not present themselves a.s exclusive. We may shift from one metaphoric orientation to another. At one time we see the computer as intelligent assistant, at another time we are overwhelmed by the dramatic possibilities of tht computer.

Such orientations do not pertain only to the early conceptualiz`~tion of a particular computer system design. Working with computers carries the added interest that a single program is usually a

collaboration by systems developers influencing one another indirectly through the environments they create. For example, the devel opers of programming languages provide environments and sets of metaphoric orientations for those who write high-level programming tools, such as HyperCard, Macromedia Director, and so on. Other developers then use these tools to create environments in which computer users can customize their own tools. Of course this is never a simple top-down process. The way programmers formulate their environments is strongly influenced by the metaphors prevalent in the user's domain. A field of metaphors pervades the various intersecting communities of developers and users.

There may also be major shifts in metaphorical emphasis, analogOtlS to Kuhnian paradigm shifts in science. Some major shifts that have filtered through the various levels of systems development at one stage or another include the computer program and the problem domains they serve conceived as a flow of control through an intricate network of decisions. The original Fortran programming langtlage required the programmer to make these control flows explicit, through "go to" statements and subroutine calls. Then there were structured programming languages, such as Pascal, which allowed for constructing the computer program and conceptualizing the problem domain in terms of autonomous subtasks or procedures understood hierarchically. These procedures exist as modules that can be assembled and linked together to accomplish higher-level tasks. There are also object-oriented programming languages, which encourage the systems designer to think in terms of autonomotls ohjects with behaviors and properties. The kit of parts becomes very flexible. The idea of object-oriented programs now pervades the design of most programs, from word processors to CAD systems. And no doubt there are other potentially pervasive metaphorical orientations in the pipeline or about to be resurrected.

So in software design, we are being encouraged to participate in the metaphors we set for each other. Some high-level programming tools ("authorware") encourage particular orientations to the computer through the metaphors they present to systems designers. For example, some programming environments for creating expert sys

295 Metaphors and Machines

tems are designed around notions of autonomous agents, group decision making around a "blackboard," "inference engines," knowledge bases, and so on. Some multimedia systems are designed around scripts, frames, and cast members_metaphors derived from animation and film production.

Metaphor play also applies at the level of software programming, writing the code or "hacking." The programmer frequently copies segments of programs from elsewhere. Segments of programs are taken from one situation and applied to a new situation. This involves "seeing as"_seeing a previous problem as this problem and this problem as a previous problem. Program components, such as variables, operators, and data structures, take on different characteristics. For example, variables are sometimes seen as containers, at other times as algebraic terms, generalizations, placcholders, pointers, or handles. The program is treated as a conduit into which the programmer pours effort, code, data, or it is seen as a defective structure that needs repairing, or an experiment that involves testing.

The power of metaphors to define problem regimes and prompt action suggests a particular approach to design practice. The architectural design practitioner does not come to a situation with fixecl, predefined problem statements but undertakes an investigation and engages in dialogue through which appropriate metaphors emergt. These metaphors are arrived at by both the practitioner and the client in the specific situation. Problems are presented and acldressed through such exchanges and collaborations. The metaphoric view of design points to design as a richly collaborativt activity. So too does software design. Clear and unambiguous prohlem statements are rare. Any problem statement needs to be interpreted in a situation. Most software design proceeds by rapicl prototyping. A computer system is proposed and developed for a particular understanding of a problem situation and tested either informally or formally. The requirements of the program emerge as the program undergoes development. The program results fron1 an interplay between what is wanted and what is possible. Therc is always a collaboration among designers, sharing and developing

metaphors as suggested above, and with clients and users, who bring their own developing understandings and metaphors to the situation.

There are, therefore, at least two extreme modes of softwaredesign practice. On the one hand, the methods orientation relies on the distinction between problem and solution and on hierarchical design procedures progressing from problem statements to performance specifications to flowcharts to informal pseudocode to operational programming to testing and evaluation. On the other hand, there is the mode of software-design practice that recognizes what every programmer knows_that you are a skilled programmer by virtue of skills and experience acquired by many years of practice, that each new project involves a new "method," and that you frequently do not know what a program should deliver until you have investigated what is possible, which is often not known until the end of the project. Of course there are contractual requirements from the client that the project will be completed on time and to a particular specification. The management of the project and the professional practice of the company (with their metaphors of timelines, milestones, and the assurance of total control) are often at odds with the metaphors of the design and development team.

So metaphor is implicated in the design of computer systems. Metaphor provides a metaphor for understanding design. It is not the only metaphor, nor is it the most perspicacious in all situations. But then to appeal to logic, flow diagrams, rewrite rules, Turing machines, and other paraphernalia of programming theory is not to abandon metaphor but to shift to a different set of metaphors. The appropriateness of any metaphor depends on the context. The pragmatic view of metaphor asks is this metaphor enabling in this situation?

Metaphor theorists have said little about design explicitly, though Johnson and Ricoeur in particular have related metaphor to the workings of the imagination. Design, whether of buildings or of sof tware, is an intensely imaginative activity. But then is this any different from other supposedly rational/logical activity?Johnson emphasizes that the workings of the imagination are not to be distinguished from rational thinking. There is a strong presupposition promoted

297 Metaphors and Machines

throughout the philosophical tradition, notably by the empiricists, that the imagination is different and subservient to rational thinking. Hume, for example, described the image as the residue of an impression, which is to imply that imagination is a secondary, derivative faculty.93 To reinstate the importance of the imagination is also one of the missions of romanticism. Though it elevates imagination, romanticism trades heavily on upholding a distinction between the imagination and reason. Pragmatism (and phenomenology) seeks to unite the two, showing that the distinction is one of word usage but not of substance. According to Dewey, thinking that is of any consequence at all pertains as much to the aesthetic and the imaginative as to what we may choose to call pure reason, whether practiced by the artist or the scientist. In keeping with the pragmatic tradition, Johnson seeks to reinstate the imagination as essential to all aspects of human cognition.

Imagination is a pervasive structuring activity by means of which we achieve coherent, patten~ed, unified representations. It is indispensable for our ability to make sense of our experience, to find it meaningful. The conclusion ought to be, therefore, that imagination is absolutely central to human rationality, that is, to our rational capacity to find sign)ficant connections, to draw inferences, and to solve problems.94

This imaginative structuring implicates metaphor. To see this thing as that thing is creative and imaginative through and through. According to Johnson, "Metaphorical projection is one fundamental means by which we project structure, make new connections, and remould our experience."95

Ricocur adds support toJohnson's view. Following Kant, he says, "we have to look at imagination as the place of nascent meanings and categories rather than as the place of fading impressions."96 For Ricocur, theories of imagination return us to the pre-Socratic tradition. Metaphor works through opposition, dialectic, and difference.

According to Ricocur, to "see sameness in difference is the genius of metaphor."97 On his view, the tension is notjust between the two terms of the metaphor (e.g., "house" and "machine" in "a house is a machine") but in the relation itself, in the copula ("is"). The tension is between sameness and difference. In metaphor, we detect sameness in spite of the overwhelming presence of difference.98

Ricocur does not advance a full theory of imagination but asserts that imagination is "the emergence of conceptual meaning through the interplay between sameness and difference."99 Metaphor makes obvious the process that is covert in all attempts to derive meaning, namely, the conflict between sameness and difference. Metaphor brings this conflict between sameness and difference to light. It does this by defying categories. Metaphor bypasses accepted categories to reveal "unnoticed similarities in the field of our experience."'__ A metaphor is a classic case of what Ryle and Turbayne characterize as a "category mistake."'_' We "misclassify" a house as a machine.

To highlight the tensional structure of metaphor echoes our earlier discussion about metaphoric versus literal language. There is a tension inherent in asserting a metaphor: is the assertion a truth, a fiction, or a metaphor? This tension is always present in language but is rendered obvious through the issue of metaphor. The identification of this tension echoes the tensional relationship between the "is" and the "is not," being and nonbeing, out,!ined in the philosophical tradition of Hegel and Heidegger. It could be said that metaphor resides at the interstice between "is" and "is not." Translating this insight into the world of information technology, every assertion about the computer is in tension with a denial of the assertion. The computer is/is not an intelligent entity. The configuration of pixels on the computer screen is/is not a sheet of paper.

This tensional, dialectical account of metaphor accords with Heidegger's notions of truth. The conventional view of truth, dating from Plato and made explicit by the empiricists and logical positivists, pertains to how assertions correspond to some state of affairs. For example, the truth status of the proposition "the computer is intelligent" is thought to reside in the degree of correspondence between the proposition and the evidence of our senses to support the claim. According to logical positivism, for a statement to be meaningful it must at least be testable. The truthfulness of "the computer is intelligent" resides in our ability to elaborate and define the terms of the expression and to establish some means of testing

299 Metaphors and Machines

or refuting it. (The Turing test is such a test. ) According to the correspondence theory of truth, the statement is meaningful but false. Similar statements subject to truth testing include the "computer is a theater," "a movie," "an inkblot," and so on. In terms of correspondence, these assertions are patently false. From Davidson's view, we may infer that although these statements are false, their meaning, which is simply their use, resides in their status as metaphors. By this account, the idea of metaphor is subservient to the idea of correspondence. Metaphor accounts for the uses we are able to make of untruthful statements.

Heidegger's notions of truth as disclosure take us in a different direction of inquiry, however, and in a different direction regarding the workings of metaphor. Whereas truth as disclosure has its place and is "correct" according to Heidegger, that is not the "essence" of truth. We use the term truth in many different contexts. We say of a painting that it is truthful. According to the correspondence theory of truth, this means that the painting corresponds to something else_a scene or a particular person. But this is clearly an impoverished view of truth in art. For Heidegger, the truth of a painting is what it discloses or reveals. Any work of art both reveals and conceals, and appreciating that truth is clearly a matter of interpretation in a context of understanding. If this is so of art, then it is equally the case in other contexts_including observations in science, the evaluation of mathematical models, and the interpretation of assertions in language. Language has a revelatory propert`,~ for Heidegger, and this is where truth resides_not in what words correspond to but in what they reveal. Heidegger thinks this view of truth is far removed from a murky relativism. Its roots are in the flux and play of strife, conflict, and tension_human rationality as understood by the pre-Socratics and appropriated by Eckhart and Hegel.

The dialectical tensions in a situation of interpretation are described in different ways by Heidegger and others_as the strife between earth and world, revealing and concealing, and so on. Ricocur is more straightforward. He uses the idea of poetical language to illustrate this tension. Poetic, or metaphoric, language changes our

way of looking at things: "What is changed by poetic language is our way of dwelling in the world. From poetry we receive a new way of being in the world, of orientating ourselves in this world."'02 How does this occur? According to Ricoeur the purpose of metaphorical language is not to improve communication nor to ensure clarity of argument, "but to shatter and to increase our sense of reality by shattering and increasing our language. The strategy of metaphor is heuristic fiction for the sake of redescribing reality. With metaphor we experience the metamorphosis of both language and reality." lo3

Elsewhere Ricoeur shows how poetry brings to light the tensions within language. The conventional distinctions between the literal and the metaphoric point to the "paradox" of the copula. This paradox preserves our sense of being- in-the-world.

Here are summed up all the forms of 'tensions' brought to light by semantics: tension between subject and predicate, between literal interpretation and metaphorical interpretation, between identity and difference.... They come to completion finally in the paradox of the copula, where being-as sign)fies being and not being. By this tu~n of expression, poetry, in combination with other modes of discourse, articulates and preserves the experience of belonging that places man in discourse and discourse in being."04

Poetry and metaphor are therefore primordial, and, in opening up the possibility of alienation, they allow us to reflect on being and dwelling: "What is given to thought in this way by the 'tensional' truth of poetry is the most primordial, most hidden dialectic_the dialectic that reigns between the experience of belonging as a whole and the power of distanciation that opens up the space of speculative thought."'05 What can be said of poetry and metaphor can be said of technology. Technology is a means of establishing difference and thereby establishing a sense of dwelling, if we let it. In this regard technology functions as metaphor. The progress and upheaval of technology is like the changes in metaphor that excite our sense of belonging.

The notion of difference has implications for the design of computer systems. We can appropriate the differences that a metaphor reveals as a stimulus to design activity. The notion of difference also

Notes to Pages 245-254

78. See A. Rand, The Fountainhead.

79. See Cunningham and Jardine, Romanticism and the Sciences.

80. The return to crah values in architecture could be regarded as romantic. See for example Charles, Prince of Wales, A Vision of Britain: A Personal View of Architecture.

81. Jones, Design Methods, 62-63.

82. Broadbent, Design in Architecture, 456-57.

83. Ibid., 322.

84. See B. Edwards, Drawing on the Right Side of the Brain; and D. Kolb, Experiential Learning: Experience as the Source of Learning and Development.

1. T. D. Erickson, "Working with interface metaphors," The Art of Human- Computer Interface Design, 65-73.

2. Ibid., 65.

3. The computer operating system is the suite of programs that controls the interface and the operations common to all programs, such as reading from, and writing to, files.

4. For a description of object oriented programming, see A. Goldberg and D. Robson, SMALLTALK-80 The Language and Its Implementation.

5. See S. Turkle, "Computer as Rorschach," Society (January/February): 15-24.

6. As l will show, principle and method can be made subject to the idea of metaphor.

7. See T. H. Nelson, "The right way to think about software design," The Art of Human-Computer Interface Design, 235-43.

8. Ibid., 236.

9. Ibid., 237.

10. Ibid., 241.

11. A. Kay, "User interface: A personal view," The Art of Human-Computer Interface Design, 191 -207.

12. Ibid., 199.

13. For an introduction to metaphor, see D. E. Cooper, Metaphor; G. Lakoff and M.

Johnson, Metaphors We Live By, A. Ortony ea., Metaphor and Thought, 1. A. Richards, Philosophy of Rhetoric, and M. Salner, "In;roduction: Metaphor and human understanding," Saybrook Review 7: 1-19. For a discussion of metaphor and models in the

359

context of design, see A. B. Snodgrass and R. D. Coyne, "Models, metaphors and the hermeneutics of designing," Design Issues 9 ( I ): 56-74.

14. Aristotle, Poetics, 1457b. See W. D. Ross, ea., The Works of Aristotle, vol. 11.

15. Aristotle, Rhetoric, 1404b.

16. According to Aristotle, there are four kinds of metaphor in poetry and rhetoric. There is metaphor as a transferral from genus to species, from spec~es to genus, and from species to species, and metaphor by analogy. See Aristotle, Poetics, 1457b.

17. J. Derrida, "White mythology: Metaphor in the text of philosophy," Margins of Philosophy, 207-71.

18. F. Nietzsche, The Complete Works of Frieduch Nietzsche, vol. 2, Early Greek Philosophy and Other Essays, 180.

19. H.-G. Gadamer, Philosophical Herfneneutics.

20. Ibid., 24.

21. P. Ricoeur, The Rule of Metaphor.

22. Ibid., 12.

23. N. Goodman, Ways of Worldmaking.

24. As explained in David Wood, Philosophy at the Limit, 29.

25. See M. Hesse, "The explanatory function of metaphor," Revolutions and Reconstructions in the Philosophy of Science, 111-124.

26. M. Black, "More about metaphor," Metaphor and Thought, 19-43.

27. See for example D. Gentner, B. Falkenhainer, and J. Skorstad, "Viewing metaphor as analogy," Analogical Reasoning: Perspectives of Artif cial Intelligence, Cognitii~e Science, and Philosophy, 171-77.

28. D. Davidson, "What metaphors mean," On Metaphor, 29-45.

29. Ibid., 30.

30. Ibid., 31.

31. Ibid., 32.

32. Ibid., 39.

33. Ibid., 40.

34. Ibid., 41.

35. Ibid., 44.

36. See also R. Rorty, "Unfamiliar noises: Hesse and Davidson on Metaphor,"Journal of the Aristotelian Society (July): 283-96; and Mary Hesse, "Unfamiliar noises_ tropical talk: The myth of the literal," Journal of the Aristotelian Society (July): 297311.

37. See G. Lakoff, Women, Fire, and Dangerous Things: Wh.at Categories Reveal about the Mind; M.Johnson, The Body in the Mind: The Bodily Basis of Meaning, Imagination, and Reason; and Lakoff and Johnson, Metaphors We Live By.

38. See M. Turner, "Categories and analogies," Anah~gical Reasoning: Perspectives of Artificial Intelligenc.e~ Cognitive Science, and Philosophy, 3-24; and E. Rosch, "Categorization of natural objects," Annual Review of Psychology, 32:89-115.

39. Lakoff, Women, Fire and Dangerous Things, 271.

40. Ibid., 267.

41. Ibid., 77.

42. See D. Schon, Displacement of C,oncepts; and D. Schon, "Generative metaphor: A perspective on problem-setting in social policy," Metaphor and 7'hought, 254-83.

43. As discussed later in this chapter, arguments about the origins of metaphor turn in on themselves. The quest for an origin indicates a particular metaphorical orientation. Grounding metaphor in bodily experience (a view criticized by Dre~fus) suggestc that there are foundations for our metaphors, a particularly Cartesian metaphorical orientation. Metaphor theorists generally recognize (and celebrate the fact) that we are never able to free ourselves from metaphorical orientations. There is no nonmetaphorical language at our disposal for explaining what metaphor is and how it works. All statements about metaphor are metaphorical. That a metaphor is the juxtaposition of two ideas is a metaphor. According to Ricoeur, there is no mastery and control of metaphor by means of class)fication. See H. L. Dreyfus, Beingin-the-World: A Commentary on Heidegger's "Bang and Time, "Division l; and P. Ricoeur, 7'he Rule of Metaphor, 18.

44. LakoffandJohnson's argument may, however, support the art)ficial- intelligence paradigm of connectionism.

45. For some alternative conceptions of the Ih1e, see G. Deleuze, The Delenze Reader, 165-72.

4fi. This includes chaotic, nonlinear, and complex systems, as well as connectionism.

47. Johnson, The Body in the Mind, 104. There have been other choices in the development of the computer that reflect other metaphorical preoccupations. Digital (numerical) processing has been favored over analog. Digital computing has been favored over neural (connectionist) computing. Raster display technology has been favored over vector display. We have generalized computer systems rather than specialized systems, distributed (personal) computing rather than mainframe (centralized). Computer drawings and models are described with orthogonal grids rather than polar coordinates. It is common to store the endpoints of lines rather than _ectors or equations. Often objects rather than actions are stored. The emphasis in drawh1g descriptions is on precision, photorealism, and now virtual reality rather

than communication or expression. In each of these choices, and others, communities of researchers, inventors, designers, and systems developers have been led h certain metaphorical orientations. These metaphors have defined what is a necessit, in a particular situation.

48. Note that such writers as Cadamer rehabilitate Aristotle's thought from such reductive interpretations. See H.-G. Gadamer, Philosophical Henneneutics.

49. See J. Derrida, Of Grammatology; J. Culler, On Deconstruction: Theory and Cnticism after Structuralism; and C. Norris, Deconstruction: Theory and Practice.

50. See J. Derrida, " Geschlecht 11: Heidegger's hand,'' Deconstruction and Philosophy: The Texts of Jacques Derrida, 161 - 196.

51 Of course the idea of usefulness is not unproblematic. I am here contrasting the pragmatic view with that of deconstruction. Note that Derrida makes no such distinctions between the pragmatic/scientific domain and that of philosophy/literature. For example, Derrida is ruthless in his extended criticism of Austin's and Searle's language theories.

52. In presenting this argument, l am echoing Derrida's identification of the kind of writing that precedes ordinary writing. This is a protowriting, of which both speech and ordinary writing are examples.

53. See M. Heidegger, Being and Time, 93; and Dreyfus, Being-ln-7'he-World, 89.

54. The counterview is that both the computer and science are products of technology.

5F1. See M. Hesse, "The explanatory function of metaphor," Revolutions and Reconstructions in the PhilosotJhy of Science, 111-24.

56. Ricoeur, The Ricoeur Reader, 84.

57. Hesse, "The explanatory function of metaphor," 1 12.

58. The Turing machh1e is named after Alan Turing. one of the founders of moder n comptlthlg, who first proposed such a universal algorithmic machine. I describe i briefly in chapter fi.

59. For a full explanation, see D. 1. A. C.ohell, Int10duction to Cormimter 7'heory.

fiO. Hesse, "The explanatory function of metaphor," 1 14.

61. Ibid.

62. Ibid., 12().

63. Ibid., 123.

362

Notes to Pages 281-289

66. See G. Ryle, The Concept of Mind; Schon, Displac.ement of GncepLs, and R. J. Sternberg, Metaphors of Mind: Conceptions of the Nature of Intelligence; E. A. Havelock, The Muse Learns to Write: Re/.7eaions on Orality and Literacy from Antiquity to the Present.

67. See M. McLuhan, The Cutenberg Calaxy: The Making of Typographic Man; W. J. Ong, Orality and Literacy: The Technologizing of the Word; E. A. Havelock, The Muse Learns to Wnte; 1. Illich and B. Sanders, ABC: The Alphabetization of the Popular Mind; and M. Heim, Electric Language: A Philosophical Study of Word Processing.

68. Ong, Orality and Literacy, 56-57.

69. Ibid., 131-132.

70. Schon, Displacement of ConcepLs.

71. For evidence of the prominence of information processing as metaphor, see C. Shannon and W. Weaver, The Mathematical Theo7y of Communication, A. Newell and H. A. Simon, Human Problem Solving; O. Akin, "How do architects design?" Artiicial Inte.lligence and Patte7n Recognition in Computer-Aided Design, 65- 119; and R. D. Coyne, M. A. Rosenman, A. D. Radford, M. Balachandran, and J. S. Gero, Knowledge-Based Design Syste7ns.

72. There is also a sense in which technologies create the problems they are designed to solve. It is often pointed out that designers and other professionals now have to cope with a great deal of information and need methods and machines to deal with this. In fact, it can be shown that the quantity of information is very much a product of the information-processing tools, rather than a need that is "objectively" recognizable and independent of the h1formation-processing metaphor. For example, remote sensing and monitoring devices generate information. Electronic mail ser_~ces and desktop publishing result in the broadcasting of vast amounts of informatim1 with little discrimination. Data, such as census results, can be manipulated in ways that produce even more information. Notwithstanding the usefulness of some of this information it can be said that the solution (computers) is also causing the problem (how do we handle all the information?).

73. Basalla, The Evolution of Technology, 107.

74. W.J. Mitchell and M. McCullough, Digital Design Media: A tlandbook fo7 Archite.cts, Landscape Architects, and Url7an Designers.

75. i leidegger's original German word, which we translate as "letting be," is GelassenheiL

76. Heidegger, Being and 7'ime, 190. 77. Ibid., 201.

8(). DrtvfIzs, Being-in-the-ll~orld, 199; Hcidegger, Beinp;and 7ime, 191.

81. M. Heidegger, The Principle of heason, 48.

82. D. Wood, Philosophy at the Limit.

83. M. Heidegger, The Prtnciple of Reason, 47.

84. Wood, Philosophy at the Limit, 35.

85. L. Wittgenstein, Philosophical Investigations.

86. Wood, Philosophy at the Limit, 35.

88. Ibid., 236-37. See also Wood, Philosophy at the Limit, 36.

89. But then, as Wood points out, there may be another, more-profound dimension to Heidegger's denial of metaphor. The concept of metaphor may be implicated in the concept of home itself. The subject matter of the particular statement in question (language as the home of being) actually bears on the matter of metaphor itself. I will return to this line of inquiry at the end of this essay.

90. This is explained in Norris, Deconstruction: Theory and Practice, 143.

91. This is explained in Wood, Philosophy at the Limit, 39.

92. See Snodgrass and Coyne, "Models, metaphors and the hermeneutics of designing."

93. P. Ricocur, The Ricoeur Reader, 82.

94. Johnson, The Body in the Mind, 168.

95. Ibid., 169.

97. Ibid., 80.

98. Ibid., 81.

99. Ibid., 82.