AI turns this interpretive investigation o f the
inanimate
around.
Brett Bourbon - 1996 - Constructing a Replacement for the Soul
Seeingnotevenasifbut through another's eyes as they see through ours is to be a more complex form of life that has at least four eyes, a network ofbrains, nerves and limbs.
The limit to this seeing is not an interpretive limit, where we use physiological (or ontological language) as a metaphor, but an ontological limit; not a limit between people or within a language game, but a limit constituting language: "Incredible the first animal that dreamed o f another animal"(Fuentes, 9).
This is what I will later call the limit between the mind and the soul.
This is the possibility to confuse the world for a mind and a mind for a world from the inside o f each.
In this sense literature becomes a philosophy o f mind, an attempt to construct and embody the mind such that human particularity and separation is reconstructed within a system of identities operating as the ideal or fundamental mind of our form o f life, approaching or enacting our species-being.
From seeing through another's eyes we are transformed into God, collapsing all time, all worlds (whose worlds? ) into our being. This might be a justification or a mad description of art. This fantasy is almost like taking language too seriously, or at least offering up the strangeness of language as so powerful as to undo itself into absurdity, as evidence o f both our extreme alienation and as the means to survive it: to find Descartes' certainty not in himselfbut in others, to allow perception or, within the context of Thoreau's allegory, language to dissolve our ontological limitations into what sound's like divine power. But why does Thoreau jump from seeing through another's eyes to seeing
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
40
? through God's eyes ("all ages o f the world in an hour, ay, in all the world o f the ages")? Is this the ontological justification for loving thy neighbor? Do we believe this fantasy? If we did then Wittgenstein could tease us back into sense by correcting our language. And yet this would be a "miracle". The ambiguity o f language, the need for philosophical therapy, protects the distance between us. Thoreau's piece of science fiction here is a version of the quest for certainty, to confirm the existence of other minds by entering into one,orevenasinartificialintelligence,bymakingone. Thoreauassumesananswerto skepticism, like Wittgenstein's notion of a true book of ethics, would re-construct the fundamental ontology of all that defines our lives and our world(s). If we admit that such a reconstruction is impossible, that such certainty is illusory, then what would the construction o f another mind amount t o ? 3
Wittgenstein answers: "If a lion could talk, we could not understand him" (223). But again, what would the construction of another mind amount to? What does 'mind' mean here? If I convert a lion into a human being I have remade my form of life: my biology, species-being, history, culture, society, conceptual, existential, and phenomenological world(s), language, desires, values, commitments. Are there necessary and sufficient conditions for being human? At what point would a lion cease to be a lion and become human?
Any attempt to answer this question, or to build a human being or a mind, or a machine X even if limited to a conceptual description (of a lion and a human; what would or should I include in such descriptions? ) or scientific models (in biology, neuroscience, cognitive science, or computer science) are stuck inside our human limitations (ontology,
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
? phenomenology, language, psychology, biology, etc. ). This means any such question or attempt can always be translated into the question 'what does it mean to be human? ' And what do things mean? A question that can be answered, at least, by science or theology. Itisasurprisethatthesetwoquestionsarerelatedorevenattimesmightseem the same.
How would we answer this question? : What does it mean to be human? Maybe withMiddlemarch, the works ofAristotle, or Bach'sMass in B minorl Ifthis question can only be answered, however poorly, with representative representations, it is not a question but a riddle. A riddle we might answer with our own life. This riddle, however, is not like the duck-rabbit 'riddle', nor like the riddle o f the Sphinx. To know the solution to the Sphinx riddle you have to know how it is the solution, If you are simply told that man is the solution to the riddle, but do not understand why this is the answer you do not understandtheriddle. Knowingtheansweristoknowhowtoapplytheriddleasa description or a picture of human life. One must know what is relevant, what aspect of human life is being pictured by four, then two, then three legs. 4
The riddle 'what does it mean to be human' is not like this. One could call it instead a riddle of the enveloping facts. Every 'answer' to the riddle is a restatement of anotherrelatedriddle. Toanswer'whatdoesitmeantobehuman? 'withBach'sMassin BMinormeanstheriddleofhumanityistheriddleofBach'sMass. Howthisisthe answer cannot be explained except as another riddle. We can never decide what is relevant aswecanwiththesphinx. Wittgensteinarguesthatonemightseesomearbitrarycipherin any number o f ways, "in various aspects according to the fiction I surround it with. And
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
42
? here there is a close kinship with 'experiencing the meaning of a word"' (PI p. 210). The fiction surrounding 'the meaning of being human' cannot reduce the riddle to anything like a word; we can never decide what is relevant as we can with the sphinx. Joyce represents the riddle of the meaning of being human as (or in) the riddle ofFinnegans Wake:
Yet to concentrate solely on the literal sense of even the psychological content of any document to the sore neglect of the enveloping facts themselves circumstancing it isjust as hurtful to sound sense (and let it be added to the truest taste as were some fellow in the act of perhaps getting an intro from another fellow turning out to be a friend in need of his, say, to a lady of the latter's aquantaince . . . straightaway to run off and vision her plump and plain in her natural altogether, preferring to close his blinkhard's eyes to the ethiquethical fact that she was, after all, wearing for the space o f the time being som definite articles o f evolutionary clothing. . . (FW109. 10-21)
This difficulty of answering this riddle forces us inside the question. We can translate it into philology with the help ofsingle quotation marks: What does it 'mean' to be a 'human being'? We can answer this by showing our uses o f'mean' and 'human being' within whatever conceptual logics they can function. Understood in this way philology becomes the analysis of the functioning of these categories, and in this an analysis of meaning. 'Function', however, is slippery. Does a crystal have a function of growing, or does it simply grow? Function implies a kind of intentionality, a directedness and a teleology. Function in this sense means to unite proximate causation (the causes describing how something works) with final causation (telos). Final causes are either
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
43
? epiphenomena (a virus makes you sick not because it is malevolent, but as an effect o f its living) or ascribed interpretations (descriptions o f motives, reasons, and justifications).
The riddle in this philological form, like our use o f 'function', constantly confuses interpretations for cause, as does Freud when he understands the unconscious, an interpretative category, to function as part o f psychic mechanism. Freud offers interpretations and he has no way of establishing that the structures he describes or 'discovers'causeanything. Causescanbetested:ifxhappens,thenyfollows(inclearly defined conditions): Freud cannot establish this kind o f connection. He describes instead motives and justifications and reasons that are interpretations of our actions and expressions. They cannot be tested, only applied.
But Freud's confusion of interpretations for psychic causes is a temptation hard to resist, especially when it can generate important questions. Inquires into the meaning of the phrase 'human being' can generate, for example, the following questions:
What rights do human beings have?
What counts as being human?
How do human beings represent themselves?
In these questions 'meaning' seems to describe the criteria for what counts as being 'human'. The criteria for being human might take the form of descriptions of kinds of humans, a set of traits or properties, or in literature the articulation of a way of being human, as a man or woman, an artist or engineer, or as a member of some ethnic or cultural group defined in whatever way. The meaning o f such descriptions is determined
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
44
? by their inclusion or articulation within some allegory, through an interpretation; in this case meaning is often equated with interpretation.
We do not know, nor within this methodology do we have any way of determining thestatusofthesecategoriesorallegories: aretheytrueordotheyevenexist. Wemight be very clever in interpreting some event or the actions or identity of someone in relation to the devil or some demonology but we do not know how to ask within philology, if the devil exists or if this allegory is an adequate explanation of someone, or if we are using this interpretation to further our own ends, the power o f the church against women for example. This may be a result of cultural blindness, but it is blindness nevertheless.
This limitation drives us further into the riddle.
Attempting to distinguish the domains of interpretation and causation force us to investigate the relation between what is real [ontological claims] and what is meaningful [semantic functions]. We retranslate the question into a kind of philosophy: W hat does it 'mean' 'to be' 'human'? 5 What does 'to be' mean? Such a question asks about the semantics of self-reflection. We can now ask what is the ontological, epistemological, and axiological (value) status ofthe criteria, interpretations, categories, and 'entities' generated in the previous versions o f the question. What it means to be human must be understood in relation to the criteria or the processes determining what is real. This 'reality' is unstable, but this instability requires that the descriptions, interpretations, and representationsofhumanbeingsfunctioninrelationandrespondtothisinstability. This entails an attempt to distinguish between interpretations and causes. Cognitive science does this by assuming that only what can be made can count as real (although humans may
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
45
? not be the 'maker'): being human must be understood as an effect. The distinctions between animate or inanimate, human and non-human (animal or thing) dissolve in the demand that any such distinctions be produced as a process outside ofthe interpretive structurewhichdefinesthem. Thisdoesnotmeanthatonecanstepoutsideofthe hermeneutic circle, but that the question about whether something is animate and inanimate must also rest on whether such a distinction exists as a product o f causes. Can knowing or recognizing the difference between what is animate or inanimate or between human and non-human mean anything separate from not only how these distinctions are used but also how these distinctions are possible?
Wittgenstein states: 'A metaphysical question is always in appearance a factual one, although the problem is a conceptual one" (Remarks on the Philosophy o f Psychology I, ? 949). Wittgenstein analyzes how we use our concepts and shows how our instantiation or reification ofthese concepts is often unwarranted. Is that the only conceptual problem? If we agree that our claims about the world (or picture of our mind as inside and the world as outside: o f our understanding as a state, or o f consciousness as states or entities) should only be understood as meaningful within our language and that these claims illegitimately instantiate some truth or reality of our being or the world, then how do we commit ourselves to these concepts as if they were real? If we determine the
grammatical use o f 'understanding' and the rules determining its use as nonsense, we have notunderstoodourinvestmentofourselvesintheconceptofunderstanding. Wittgenstein is looking at the points where we mistake our grammar for ontology. As Hacker argues the claim that 'Nothing can be red and green all over' is not a truth about the world. This
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
? does not mean that it is false. We cannot say what it would be like for it to be true or false. It is nonsensical because the grammatical rules determining the sense o f something being only red and being only green do not allow for something to be only red and only green simultaneously(//mg/rt 197). 6 Sense describes the possibilities of our language; nonsense describes what has been excluded from our language (PI? 500).
The instantiation o f distinctions that function within language games means that we have a specific problem in determining the status o f these distinctions when they are used to describe ourselves or to instantiate the meaning or cause or functioning of behavior, language, attitudes, beliefs, and so on. All psychological, intentional, ideological language has an unknown and problematic existential status. How do we invest ourselves in our self-descriptive language games? This is not to ask 'Are we conscious? ' or 'Do we have belief states? ', etc. If the logic of language is such that the distinctions formalized as words, or sentences, or concepts function and have meaning only within (a) system(s) of discourse, that is, in relation to other words, sentences, concepts o f a particular language game, then even our interpretative distinctions and models are only a shuffling of tokens
or the working out ofthe logic organizing a particular language game. A working out which has no truth values separate from the rules o f application and criteria o f sense that
constitute the language game as a coherent system o f allegory that people use bound by some conventions. So much o f literary criticism and philosophical analysis devolves into the shuffling oftokens within these systems ofthe extension ofthe number oftargets for allegorical assimilation within a discourse o f a community o f speakers.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
47
? The problem of how we enter into our interpretations or how we are involved in our seeing determines much of Wittgenstein's analysis in the latter part of part I and all of part II of the Investigations. Wittgenstein's description of our involvement in and between language games (in our negotiation between sense and nonsense) describes a form of temporality as one o f the limits o f our mind. Wittgenstein understands this limit to be described by the shifting relationship between what he calls our language games and our form o f life: "Here the term 'language game' is meant to bring into prominence the fact
thatthespeakingoflanguageispartofanactivity,orofaformoflife"(PI? 23). Meaning can emerge through our involvement within a language game constituted by the limits of our form of life. In Finnegans Wake this same relationship is described within a different kind of obscurity. This is a description ofFinnegans Wake, "this daybook, what curios of signs (please stoop), in this allaphbed! ":
Can you rede (since We and Thou had it out already) its world? Its the same told of all. Many. Miscegenations on miscegenations. Tieckle. They lived und laughed ant loved end left . . . In the ignorance that implies impression that knits knowledge that finds the nameform that whets the wits that convey contacts that sweeten sensation that drives desire that adheres to attachment that dogs death that bitches birth that entails the ensuance of existentiality. (FW18. 18-28)
To be before the world ofthe book is to have had the book before "The soul of everyelsesbody rolled into its olesoleself' (FW329. 18-19) in continuous reinterpretation (into categories o f miscegenation's) which alter the shape (or color) o f the world.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
48
? How do we draw a line between the animate and the inanimate? The decisions about where and how to draw this line would seem to determine who we count as like us, whoever we are, and thus it might describe our moral categories, who is good or bad; Or where and how we draw this line might explain or describe how we interpret the world and ourselves, determining our social orders or psychological identity. This means that when I ask myself what counts as human I begin circling around a (the) confusion between citizenship and personhood. This confusion is acknowledged in our use of the concept of rights. If I recognize someone's rights (as a human being), do I recognize someone as human? Didn't I recognize her human form, or did I imagine she was a machine? My acknowledgmentmeansthatIincludeherwithinmycommunity. Thisinclusioniswhat wecallcitizenship. Myacknowledgmentofsomeoneelseashumanonlyservestoinclude
that person as an example of my kind of humanness determined by his or her inclusion within the community I recognize myselfas functioning in. Thus, the American Bill of Rights acknowledges that rights must be given, and, therefore, that they are not human rights,butrightsofcitizenship. Andtheapplicationoftheserightsisafunctionof interpretation.
The confusion between citizenship and personhood might be necessary for any democracy. Thisconfusion,however,canonlybeansweredbyinterpretations, interpretations about origins, identity, values, marks, etc. ; but the fact that these categories are slippery makes any interpretation (positive or negative) unstable, and thus induces further anxiety about what constitutes humanness.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
49
? As human beings, our questions or motives in determining the meaning of a thing arise because o f the claim things have on us, a claim that describes the material world as somekindoflimit. Thisclaimisnothingnew. FrankensteinhaslittletodowithAIor science, except as an expression of anxiety about our interpretations of the instability of the distinction between inanimate and animate, between human and otherness. It is True that science explicitly questions this distinction, claiming that it is not given but is made, a product of a set of causes and effects that we now call evolution. But the book is not questioning these limits in relation to how they are constituted within nature (or out). Rather it examines the moral implications o f drawing the difference between human and others superficially, examining how we interpret our obligations to others. The problem o f alienation Frankenstein explores is possible because we can identify with the monster and the monster with us: and in this it is a novel asking about the relationship between
personhood and citizenship. It may suggest that our recognition as human (our inclusion in a community) must arise from a recognition of certain origins, but we are not made like the monster and yet he may describe our condition. The novel does not ask but leaves us with the question 'what does it mean to be made'? That question was partially answered and given a special claim on us by Darwin. But how would Darwin fit into a discussion about what counts as human? . His thinking explains, and is now used to explain, how things become living, how living creatures become intelligent, even become human. Do we imagine Darwin would pressure our texts away from the questions o f interpretation exemplified in the confusion between citizenship and personhood? I have to leave this question hanging, but the difficulty of fitting Darwinian thought (as opposed to rhetoric) is
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
? exactly the difficulty o f making the relation and difference between cause (what science describes) and interpretation (what literature demands) clear. This is the difficulty of constructing an intelligence in a machine.
For example how would our claims about the limits or status of subjectivity or our ideological judgments survive philosopher Daniel Dennett's description o f the necessity for selfishness within evolution? :
"As soon as something gets into the business of self-preservation, boundaries become important, for ifyou are setting out to preserve yourself you don't want to squander effort trying to preserve the whole world: you draw the line" {Consciousness Explained, 174)
The interpretations that function within, as and between the confusion between citizenship and personhood in Frankenstein mean we ask for the distinction between the animate and the inanimate, as it were, from inside the security o f acting as animate creatures.
AI turns this interpretive investigation o f the inanimate around. It investigates the difference between the inanimate and animate (or conscious and mechanical) as it were from the outside, not from the assumption that this distinction is meaningful but that it is caused: human beings or animals or plants are effects. Cognitive Science asks 'how can onegeneratequalitativeexperiencefromwhatcanonlybedescribedquantitatively? Itis, for example, much easier to teach a computer how to do calculus, than it is to teach it what a chair is. How can we get from our interpretations of ourselves to the causal mechanisms of our minds? This is, however, as much a question about and of poetry as about AI.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
51
? Keats' ends the "Ode on a Grecian Urn" with the equation: "'Beauty is truth, truthbeauty. '" Evenifonethinkthisisafailedline,theinterrogationofthemeaningofa thing, o f the urn in Keats' ode, suggests that any other answer would fail to give the thing meaning. Even this answer gives the proper form without the necessary force that would make it true (although it might seem beautiful or it might not). But that this question, 'what is the meaning o f a thing? ', should be important to us organizes one o f the currents in the following pages, and at the very least describes a problem romantic poetry found itselffacedwith. Theequationbeauty=truthdescribes,Ithink,arequirementfororan ideal (of whatever status) of poetry in general, although how to construe truth and beauty here is always at issue. Poetry always seems confined by an asymmetry: it can enact meaning or beauty but it's claims on truth are tenuous, becoming more tenuous within modem scientific culture. Things or objects can function as the grounds for truth, and one can argue that this is always the case, but it is certainly the case when what is real is primarily described as material. Within the Keats' ode, the status of the um is complex and shifting (within the poets interrogation), but it functions at least as the vanishing point ofthereal. Thus,theasymmetryofpoetryispartlycounteredbythefactthatitistheum that speaks the equation between beauty and truth: in a rough allegory the thing, the um, with its status as true (or real) asserts an equation between beauty and truth that can only function aesthetically, as an expression of beauty. The um as thing (the real) articulates the truth as an aesthetic. The attempt to enact the qualitative (meaning, beauty, intention, and so on, what AI calls qualia, what it feels like to be conscious or to be human) through an embodiment in language describes what Coleridge called a philosophic poem: to enact
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
52
? the beautiful (or meaningful or intentional) as real and true: poetry with ontological force, the force ofthe real which we describe as true. This is close to what is attempted in AI: to enact the equation between beauty (or the meaningful) and truth, and in this translate our interpretative relations with causal processes.
Traditional AI makes two founding assumptions about this project: what is called the symbolic system hypothesis and the algorithmic assumption. The first claims that the processes underlying intelligence are not just described symbolically but that they are symbolic: we ourselves as well as our thinking are constituted as nothing more than the self-reflexive (to varying degrees) processing of different kinds of symbols within a vast system of symbolic relations. The second claims that our mental states are algorithms, procedures that convert inputs into outputs according to a defined set of rules: a recipe or a function converting or computing inputs into outputs.
What kind of claim do these algorithms and the symbolic processes they enact have on us? AI assumes they are both true and are the only way anything can be made meaningful. In this picture, only symbolic processes, which are not interpretations, allegories or non-symbolic computations, can have sense, and thus our brains must enact these symbolic processes as our mind (There are ways even of construing all of reality as such symbolic processes within physics).
Heidegger wants to displace the claim of this picture of the world on us. Although I do not think Heidegger understood it in this way, the best sense I can make of his claim, in "Das Ding", that "the atom bomb and its explosion are the mere final emission of what has long since taken place, has already happened" ("The Thing", 166) is that the
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
53
? possibilities determined by thinking, knowledge, technology, and economics that produced a working bomb alters the kind of metaphysical possibilities in and through which we function. We can literally enact our death through our knowledge and its potential to destroy the world: I call this a metaphysical fact because our knowledge and knowing, when instituted within social power, has ontological force so great as to destroy us. The worldcanbecontainedbyourknowledgebecauseitcanbedestroyedbyit. Theatomic bomb as a metaphysical fact is a new conceptual limit determining the scope o f our commitments toward what we count as real. We can only believe in ghosts to our greater peril. Thisfact,describingthelimitoftheworldbyourknowledge,enactedthroughthe possibility of nuclear destruction, is that what counts as matter, and what matters and counts for us will have to include a commitment to a physics that can manipulate the substance o f the universe through a complex set o f mathematical models. This is called the possibility o f losing my life by losing the world. That this loss o f the world can be enacted
through human knowing, as if from a single source, gives knowing ontological force not unlike turning the world into a fantasy. What Joyce described as "lethelulled between explosion and reexplosion . . . from grosskopp to megapod, embalmed, o f grand age, rich indeathanticipated"(FW78. 04-06)hasgainedanewallegoricaltarget: wecanliterally enact our death through our knowledge o f the world as a whole and its potential to destroy the world. The problem ofthe machine is not only about our knowledge of how to be mistaken for a thing or how to be mistaken for being human. Our understanding of ourselves and the world must be a kind of reverse evolution, undoing our feelings and interpretations into quantities within some created conceptual language, a language not
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
54
? restricted by the limits o f cause or interpretation. The failure to find such a language has been the failure o f AI.
2. 3 Spiraling from interpretation to cause
The fundamental assumption for Cognitive Science is that the brain is a computer. All forms o f meaning and understanding must be derived from the computations o f neurons. 7 This Computational model grounds all mental functioning in rationality. But this rationality should not be mistaken for what we often oppose by the word
'irrationality' (these words are used primarily as interpretations). Cognitive Philosopher John Pollock defines rationality as the ordered, non-random transition between mental states: "try to imagine a creature possessing mental states but entirely unconstrained by considerationsofrationality"(HowtoBuildaPerson,70) Suchrandomtransitionswould resultinasystemscrashonanycognitivelevelonwhichtheyoccurred. Thisuseof rationality is really a way of arguing that all mental processes (metal states and their
transitions), including language instantiate an ordered system.
Pollock is attempting to construct this kind of system. He tells a story about how
to move from things and causes into interpretations and qualities; this describes an aesthetic,butnotanaestheticlikethatinanovel. ThisisthestoryofOscar. (Iwillgive only the schematics o f this story without concerning myself here with the justification o f the type/token mentalism for which he argues, although such a theory can also be understood to form an aesthetic)
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
55
? Oscar I is a system that consists o f sensory input, some information processing mechanism; which produces a picture ofthe world. This picture is generated from generalizations (computational processes) about this sensory input. These cognitive abilities are embedded in a conative (from Latin conatus, endeavor, attempt) function determining a set o f goals. Such a mental model, although Pollock ignores this, already figures symbolically a kind o f self but in this case one determined only by its pain sensors and its response. If pain then flight, if flight fails, then fight. This describes the mental functioning o f an amoebae, and therefore has its limits.
Oscar H
TheprimaryfailureofOscarIisitsvictimizationbyitsenvironment. Survival
potential increases with the ability to make predictions about the world (or really about itself about its responses to its sensory inputs). Such predictions mean that Oscar II, in replacing Oscar I, must be able to make generalizations about the environment in relation to his goals (in this case only avoid pain). This is the beginning o f mental fiction.
These predictions are generated through self-reflection: an internal sensor senses whenandwhy:pointythingsmeanpain. Isay'meanpain'becauseitoperatesasa
judgment based on a correlation between a rudimentary self (that defined by pain sensors) and the representation of pointy things. This correlation is possible by virtue of the goal (avoid pain) and the possibility for Oscar o f modeling his own sensory outputs. What these sensors do is construct generalizations about certain aspects of the environment in relationtoOscarsownresponses(pain). Suchgeneralizations,reallypossibilities,are
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
56
? logical fictions. There is no question whether they are true or false: everything that looks pointycausespain. porOscarIItheworldisfullofpaincausingthingstowhichhe always reacts as if the possibilities, the generalizations about how the world effects Oscars pain sensors, are always true (but in fact Oscar has no notion o f truth because there exists nopossibility,forhim,ofthesepredictionsbeingfalse,eveniftheyarefalse). Possibility means always true fiction. This is the mental world o f a worm.
Oscar IU
The problem with being a worm is that the environment can be deceptive and one's own sensors can fail or produce false information. Not all pointy things are alike, a pointy leaf is not like a pointy stick, and a machine-eating tiger is not the same as the mirror image of such a tiger.
Oscar HI experiences a mirror stage. Initially Oscar I and H were like amoebae to worm; Oscar n and m , however, are like bird to cat (I guess the bird has eaten the worm). The cat is able to distinguish a mirror image from another animal (it's not clear whether the cat thinks it's its own image). The bird responds to a mirror image as if it is another bird (it attacks the image)
Oscar HI must be able to evaluate the ontological status of its perceptions, are they real or not, it must be able to recognize mirrors as mirrors, and mirror-images as mirror- images. Oscar n sees the world phenomenological, as if its eyes were video or film cameras. If you hold a camera and film while walking around the picture jumps, and bounces and shakes. The viewer gets a headache trying to stabilize the picture into sense.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
57
? Why don't we see like that? We walk around and yet our picture of the world doesn't really jump around like this; why?
To answer this we must ask how can we make Oscar m generate a world separate from his senses, such that he can evaluate his sensory inputs in relation to this stable world? Pollock ask: how can Oscar distinguish between machine-eating tigers and mirror-image tigers? Asking this question means asking the difference between appearance and reality, and its answer requires an interpretation.
This level o f interpretation requires another layer o f internal sensors, this time evaluating the first level o f internal sensors. Pollock thinks that our picture o f the world is generated by two different kinds ofgeneralizations: "On the one hand Oscar HI has generalizations about the relations between his perceptual inputs and the states o f his environment, and on the other he had has generalizations about regularities within the environment that persist independently of perception of the environment"(4). This independent world is generated as a set of generalizations not about the world but about Oscars correlation o f himself (defined by avoidance o f pain) to representations (processed sensoryinput). Thiscorrelationisevaluatedinordertodeterminehigh-levelregularities that constitute a stable world, that in essence allows us to see without all the bouncing.
Now these kinds of evaluations take place at many different computational levels within the brain, and most o f these evaluations take place through hard-wired computational processes about which we have no awareness.
Pollock argues that there is a distinction between sensation and the feel of sensation. Thefeelofsensationistheoutputofsecondlevelinternalsensors,sensorsthat
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
58
? sense the first level internal sensors that process our perceptual information: being-aware means and is the sensing o f perceptual input (this is the link AI wants to make between 'mean'and'is'). Thequalitativefeelofourconsciousnessorsensingmeansandisthe processing or sensing of lower level internal sensors by higher level sensors. Consciousness and the qualitative feel o f our experience, what understanding feels like, or meaning, or love, or whatever are only what we call this second level sensory processing of our own internal processing.
I think this implies for example that all o f our emotions, love, hate, desire, fear, etc. are only interpretations of physical states. We try to make meaningful (sometimes by connecting it to some putative cause) physical states that we sense but do not understand. Even our emotions are interpretive fictions, convenient descriptions that may mask a mechanism whose primary purpose may be to simplify decision processes (emotions force us into decisions). *
What exactly does it mean to process the information from our first level internal sensor? Pollockshiftstheproblemofconsciousnessorqualitativefeelingintosome magical processing. How do these interpretations produce new effects? Or how do we invest ourselves in all o f the complicated ways we do in our correlation between internal sensors and our picture ofthe world? I don't think we should find Pollock's picture satisfactory.
Whatever the merits ofthis model ofthe mind (and there are many), it leaves us with a puzzle. So far I have described only two kinds of meta-languages, which seem to describe the limits o f the other, but whose relation remains unclear: a language o f cause
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
59
? described vaguely here by different versions of Oscar, call this Darwinian engineering, and a language o f interpretation, commonly used in literary analysis, asking about meaning, anddeterminingthedistinctionsbetweenanimateandinanimate,etc. Butthisinterpretive language undermines its own status by showing how its distinctions and categories are so unstable as to be senseless without ontological limits o f exactly the sort described by science. But these scientific limits are not foundations precisely because we cannot get outside o f our interpretations. Thus these languages o f science and literature speak in two directionsatthesamething. Ithinkthisisourinheritanceandtheforgingofanother language that includes both of these languages is the underlying although unrealized promise o f AI, and I think our problem.
1Translated and found in Kenny, "Wittgenstein on the Nature of Philosophy", 14.
2 In Being and Time.
3 Is Thoreau's jump from anyone's embodied, particular eyes to God's eyes justified by the same trick that Heidegger uses to move from Dasein-as-mine to its identity as Being-with in Chapter IV of Being and Time! The kind of holistic, non-formalizable link between particular and universal that Heidegger constructs both elucidates a critical dependence between meaning and existence, while at the same time offeringthegreatestresistancetothepossibilityofembodyingaconstructedmind. Daseinbecomesa soul. Ibelieve,however,thatHeidegger's"soul''isfalse. (Howisitpossibletotalkaboutfalsesouls? ) But I am nevertheless left with investigating the intransigent existence and meaning of our use of, belief in,andneedforwhatHeideggerpointsto. Ifconceptsarenotthings,thenwhatarethey: howisit possible for us to function at a level o f abstraction beyond materiality?
4 Cora Diamond describes what it means to understand the answer to a Sphiax-like riddle:
[T]o know the solution to a riddle is not merely to know o f something that thought o f in some way or other it is the solution: you have to know how it is the solution, If you are simply told that man is the solution to the Sphinx's riddle, but do not understand why--ifi that is, you do not understnad why man is supposed to have four legs in the morning and so on--you are in the positoion of someone who 'knows' that every equation has a solution without any idea o fhow that may be arrived at (270)
5This is, of course, not the end of the opportunities for analysis.
6 P. M. S. Hacker. Insight and Illusion: Themes in the Philosophy o f Wittgenstein. 1Neuronscanbeintwostates:excitatoryorinhibatory;onoroff:thustheyaredigitalcircuits, by stimulation a neuron reaches an activation potential and fires. The energy of this firing is not variable-- so no information can be encoded in the strength of the output--only by the fact of the firing and in, what doeschange,therateoffiring. Greaterstimulationofaneuronincreasestherateoffiring. Thereissome confusionwhetherthebraindescribedinthiswayislikeadigitalorananologcomputer, digitalswitches (or in computer science these are called logic gates) are like digital clocks, they are in one state or
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
60
? another- analog switches are like a clock with two hands--the changes in rate describe an analog relation. Itdoesn'tmatterforus--buttheassumptionbymostAItheoristsandcognitivescientistsisthat it is digital. .
8 This might be a way o f explaining how replicants in the movie Blade Runner can generate emotions even though they are designed not to have any.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
61
? 3
Romantic Fragments and Modernist Machines
InalettertohissisterandbrotherKeatsimaginessoulsasmadenotgiven. This making proceeds through what is given: the world as circumstance and the mind as intelligenceandmemoryandemotion. Whywouldapictureofthemindexpressa theology or a moral stance? "The point at which Man may arrive is as far as the parallel state in inanimate nature and not further" (326). 1The inanimate is the limit o f human happiness. Thislimitengenderstheworldasa"valeofSoul-making",rewritingthe Biblical "a vale o f tears. " In support for this inanimate limit, Keats animates a rose with sensation, in order to tell a tale of crucifixion:
For instance suppose a rose to have sensation, it blooms on a beautiful
morning, it enjoys itself, but then comes a cold wind, a hot sun--it cannot escape it, it cannot destroy its annoyances--they are as native to the world as itself--no more can man be happy in spite, the worldly elements will prey upon his
nature. (326)
The unity between the divine and human, between fundamental ontological difference, is twisted toward biology, into a unity between rose and human. In order for Keats to tell the story of our inhabitation and involvement in the world as a limitation on our happiness he endows a rose with a conscious o f that which threatens not its consciousness specifically, butitslife. Sucharoseisnotnativetoanyrealworld,butevenitsfantasticsentience
Notes for this chapter are on page 102
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
62
? does not free it from its containment and belonging to a world described by power, not determined by the roses intentions.
The rose slips from the perception ofsensation to an interpretation ofthe meaning ofthese sensations as 'annoyances'. Even to evaluate negative stimuli as annoyances implies a complex o f responses that involves the possibilities o f happiness and hope, a language ofintention in which the consciousness or the desires or beliefs or values ofthe rosestructureitsrelationtotheworld. Therosefunctionsasarosewithinasystemof opposedintentions. Intention,inthiscaseannoyanceorhappiness,isafunctionofthe systematic relation between the rose and the world, and is therefore in neither. The world is animated and the intentionality o f our involvement in the world is determined and guaranteedbythisanimation. Coleridgepicturesthisin"TheEolianHarp":
And what if all of animated nature
Be but organic Harps diversely framed,
That tremble into thought, as o'er them sweeps Plastic and vast, one intellectual breeze,
At once the Soul of each, and God of all?
Thought is generated as a consequence of our involvement with dominant non-human forces;intentionfunctionswithinthesystemofworldandbeings. Ourintentioncanbe generated by reducing ourselves to things--and then displacing intention into the environmentthroughwhichthisintentionbecomesvisible. Weenterintothedominionof arbitrary forces to the degree that we can let the rose stand for ourselves, allegorize ourselvesasusedbytheworld. Inthissensewearealwaysinvisibleinthisallegory. We
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
63
? exist as nothing more than our own interpretations o f ourselves through our language or actions. Allegories describe and constitute a limit to our knowing ourselves.
Consequently, it is a mistake to imagine that these allegories can describe our becominganything. Theycanonlydescribehowourinterpretationschange. Whenthey describe change they describe it teleologically, as if the allegory were a machine. Even if wearemachineswefunctiononlypartiallythroughourinterpretations. 'Causation'hasa particular grammar that overlaps but is not continuous with 'interpretation'. Wittgenstein at one point investigates how the word 'reading' describes how we read words into speech, how words 'cause' our reading, not our understanding o f what we read.
But why do you say that we felt a causal connection? Causation is surely something established by experiments, by observing a regular concomitance of eventsforexample. SohowcouldIsaythatIfeltsomethingwhichisestablished by experiment?
From seeing through another's eyes we are transformed into God, collapsing all time, all worlds (whose worlds? ) into our being. This might be a justification or a mad description of art. This fantasy is almost like taking language too seriously, or at least offering up the strangeness of language as so powerful as to undo itself into absurdity, as evidence o f both our extreme alienation and as the means to survive it: to find Descartes' certainty not in himselfbut in others, to allow perception or, within the context of Thoreau's allegory, language to dissolve our ontological limitations into what sound's like divine power. But why does Thoreau jump from seeing through another's eyes to seeing
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
40
? through God's eyes ("all ages o f the world in an hour, ay, in all the world o f the ages")? Is this the ontological justification for loving thy neighbor? Do we believe this fantasy? If we did then Wittgenstein could tease us back into sense by correcting our language. And yet this would be a "miracle". The ambiguity o f language, the need for philosophical therapy, protects the distance between us. Thoreau's piece of science fiction here is a version of the quest for certainty, to confirm the existence of other minds by entering into one,orevenasinartificialintelligence,bymakingone. Thoreauassumesananswerto skepticism, like Wittgenstein's notion of a true book of ethics, would re-construct the fundamental ontology of all that defines our lives and our world(s). If we admit that such a reconstruction is impossible, that such certainty is illusory, then what would the construction o f another mind amount t o ? 3
Wittgenstein answers: "If a lion could talk, we could not understand him" (223). But again, what would the construction of another mind amount to? What does 'mind' mean here? If I convert a lion into a human being I have remade my form of life: my biology, species-being, history, culture, society, conceptual, existential, and phenomenological world(s), language, desires, values, commitments. Are there necessary and sufficient conditions for being human? At what point would a lion cease to be a lion and become human?
Any attempt to answer this question, or to build a human being or a mind, or a machine X even if limited to a conceptual description (of a lion and a human; what would or should I include in such descriptions? ) or scientific models (in biology, neuroscience, cognitive science, or computer science) are stuck inside our human limitations (ontology,
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
? phenomenology, language, psychology, biology, etc. ). This means any such question or attempt can always be translated into the question 'what does it mean to be human? ' And what do things mean? A question that can be answered, at least, by science or theology. Itisasurprisethatthesetwoquestionsarerelatedorevenattimesmightseem the same.
How would we answer this question? : What does it mean to be human? Maybe withMiddlemarch, the works ofAristotle, or Bach'sMass in B minorl Ifthis question can only be answered, however poorly, with representative representations, it is not a question but a riddle. A riddle we might answer with our own life. This riddle, however, is not like the duck-rabbit 'riddle', nor like the riddle o f the Sphinx. To know the solution to the Sphinx riddle you have to know how it is the solution, If you are simply told that man is the solution to the riddle, but do not understand why this is the answer you do not understandtheriddle. Knowingtheansweristoknowhowtoapplytheriddleasa description or a picture of human life. One must know what is relevant, what aspect of human life is being pictured by four, then two, then three legs. 4
The riddle 'what does it mean to be human' is not like this. One could call it instead a riddle of the enveloping facts. Every 'answer' to the riddle is a restatement of anotherrelatedriddle. Toanswer'whatdoesitmeantobehuman? 'withBach'sMassin BMinormeanstheriddleofhumanityistheriddleofBach'sMass. Howthisisthe answer cannot be explained except as another riddle. We can never decide what is relevant aswecanwiththesphinx. Wittgensteinarguesthatonemightseesomearbitrarycipherin any number o f ways, "in various aspects according to the fiction I surround it with. And
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
42
? here there is a close kinship with 'experiencing the meaning of a word"' (PI p. 210). The fiction surrounding 'the meaning of being human' cannot reduce the riddle to anything like a word; we can never decide what is relevant as we can with the sphinx. Joyce represents the riddle of the meaning of being human as (or in) the riddle ofFinnegans Wake:
Yet to concentrate solely on the literal sense of even the psychological content of any document to the sore neglect of the enveloping facts themselves circumstancing it isjust as hurtful to sound sense (and let it be added to the truest taste as were some fellow in the act of perhaps getting an intro from another fellow turning out to be a friend in need of his, say, to a lady of the latter's aquantaince . . . straightaway to run off and vision her plump and plain in her natural altogether, preferring to close his blinkhard's eyes to the ethiquethical fact that she was, after all, wearing for the space o f the time being som definite articles o f evolutionary clothing. . . (FW109. 10-21)
This difficulty of answering this riddle forces us inside the question. We can translate it into philology with the help ofsingle quotation marks: What does it 'mean' to be a 'human being'? We can answer this by showing our uses o f'mean' and 'human being' within whatever conceptual logics they can function. Understood in this way philology becomes the analysis of the functioning of these categories, and in this an analysis of meaning. 'Function', however, is slippery. Does a crystal have a function of growing, or does it simply grow? Function implies a kind of intentionality, a directedness and a teleology. Function in this sense means to unite proximate causation (the causes describing how something works) with final causation (telos). Final causes are either
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
43
? epiphenomena (a virus makes you sick not because it is malevolent, but as an effect o f its living) or ascribed interpretations (descriptions o f motives, reasons, and justifications).
The riddle in this philological form, like our use o f 'function', constantly confuses interpretations for cause, as does Freud when he understands the unconscious, an interpretative category, to function as part o f psychic mechanism. Freud offers interpretations and he has no way of establishing that the structures he describes or 'discovers'causeanything. Causescanbetested:ifxhappens,thenyfollows(inclearly defined conditions): Freud cannot establish this kind o f connection. He describes instead motives and justifications and reasons that are interpretations of our actions and expressions. They cannot be tested, only applied.
But Freud's confusion of interpretations for psychic causes is a temptation hard to resist, especially when it can generate important questions. Inquires into the meaning of the phrase 'human being' can generate, for example, the following questions:
What rights do human beings have?
What counts as being human?
How do human beings represent themselves?
In these questions 'meaning' seems to describe the criteria for what counts as being 'human'. The criteria for being human might take the form of descriptions of kinds of humans, a set of traits or properties, or in literature the articulation of a way of being human, as a man or woman, an artist or engineer, or as a member of some ethnic or cultural group defined in whatever way. The meaning o f such descriptions is determined
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
44
? by their inclusion or articulation within some allegory, through an interpretation; in this case meaning is often equated with interpretation.
We do not know, nor within this methodology do we have any way of determining thestatusofthesecategoriesorallegories: aretheytrueordotheyevenexist. Wemight be very clever in interpreting some event or the actions or identity of someone in relation to the devil or some demonology but we do not know how to ask within philology, if the devil exists or if this allegory is an adequate explanation of someone, or if we are using this interpretation to further our own ends, the power o f the church against women for example. This may be a result of cultural blindness, but it is blindness nevertheless.
This limitation drives us further into the riddle.
Attempting to distinguish the domains of interpretation and causation force us to investigate the relation between what is real [ontological claims] and what is meaningful [semantic functions]. We retranslate the question into a kind of philosophy: W hat does it 'mean' 'to be' 'human'? 5 What does 'to be' mean? Such a question asks about the semantics of self-reflection. We can now ask what is the ontological, epistemological, and axiological (value) status ofthe criteria, interpretations, categories, and 'entities' generated in the previous versions o f the question. What it means to be human must be understood in relation to the criteria or the processes determining what is real. This 'reality' is unstable, but this instability requires that the descriptions, interpretations, and representationsofhumanbeingsfunctioninrelationandrespondtothisinstability. This entails an attempt to distinguish between interpretations and causes. Cognitive science does this by assuming that only what can be made can count as real (although humans may
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
45
? not be the 'maker'): being human must be understood as an effect. The distinctions between animate or inanimate, human and non-human (animal or thing) dissolve in the demand that any such distinctions be produced as a process outside ofthe interpretive structurewhichdefinesthem. Thisdoesnotmeanthatonecanstepoutsideofthe hermeneutic circle, but that the question about whether something is animate and inanimate must also rest on whether such a distinction exists as a product o f causes. Can knowing or recognizing the difference between what is animate or inanimate or between human and non-human mean anything separate from not only how these distinctions are used but also how these distinctions are possible?
Wittgenstein states: 'A metaphysical question is always in appearance a factual one, although the problem is a conceptual one" (Remarks on the Philosophy o f Psychology I, ? 949). Wittgenstein analyzes how we use our concepts and shows how our instantiation or reification ofthese concepts is often unwarranted. Is that the only conceptual problem? If we agree that our claims about the world (or picture of our mind as inside and the world as outside: o f our understanding as a state, or o f consciousness as states or entities) should only be understood as meaningful within our language and that these claims illegitimately instantiate some truth or reality of our being or the world, then how do we commit ourselves to these concepts as if they were real? If we determine the
grammatical use o f 'understanding' and the rules determining its use as nonsense, we have notunderstoodourinvestmentofourselvesintheconceptofunderstanding. Wittgenstein is looking at the points where we mistake our grammar for ontology. As Hacker argues the claim that 'Nothing can be red and green all over' is not a truth about the world. This
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
? does not mean that it is false. We cannot say what it would be like for it to be true or false. It is nonsensical because the grammatical rules determining the sense o f something being only red and being only green do not allow for something to be only red and only green simultaneously(//mg/rt 197). 6 Sense describes the possibilities of our language; nonsense describes what has been excluded from our language (PI? 500).
The instantiation o f distinctions that function within language games means that we have a specific problem in determining the status o f these distinctions when they are used to describe ourselves or to instantiate the meaning or cause or functioning of behavior, language, attitudes, beliefs, and so on. All psychological, intentional, ideological language has an unknown and problematic existential status. How do we invest ourselves in our self-descriptive language games? This is not to ask 'Are we conscious? ' or 'Do we have belief states? ', etc. If the logic of language is such that the distinctions formalized as words, or sentences, or concepts function and have meaning only within (a) system(s) of discourse, that is, in relation to other words, sentences, concepts o f a particular language game, then even our interpretative distinctions and models are only a shuffling of tokens
or the working out ofthe logic organizing a particular language game. A working out which has no truth values separate from the rules o f application and criteria o f sense that
constitute the language game as a coherent system o f allegory that people use bound by some conventions. So much o f literary criticism and philosophical analysis devolves into the shuffling oftokens within these systems ofthe extension ofthe number oftargets for allegorical assimilation within a discourse o f a community o f speakers.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
47
? The problem of how we enter into our interpretations or how we are involved in our seeing determines much of Wittgenstein's analysis in the latter part of part I and all of part II of the Investigations. Wittgenstein's description of our involvement in and between language games (in our negotiation between sense and nonsense) describes a form of temporality as one o f the limits o f our mind. Wittgenstein understands this limit to be described by the shifting relationship between what he calls our language games and our form o f life: "Here the term 'language game' is meant to bring into prominence the fact
thatthespeakingoflanguageispartofanactivity,orofaformoflife"(PI? 23). Meaning can emerge through our involvement within a language game constituted by the limits of our form of life. In Finnegans Wake this same relationship is described within a different kind of obscurity. This is a description ofFinnegans Wake, "this daybook, what curios of signs (please stoop), in this allaphbed! ":
Can you rede (since We and Thou had it out already) its world? Its the same told of all. Many. Miscegenations on miscegenations. Tieckle. They lived und laughed ant loved end left . . . In the ignorance that implies impression that knits knowledge that finds the nameform that whets the wits that convey contacts that sweeten sensation that drives desire that adheres to attachment that dogs death that bitches birth that entails the ensuance of existentiality. (FW18. 18-28)
To be before the world ofthe book is to have had the book before "The soul of everyelsesbody rolled into its olesoleself' (FW329. 18-19) in continuous reinterpretation (into categories o f miscegenation's) which alter the shape (or color) o f the world.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
48
? How do we draw a line between the animate and the inanimate? The decisions about where and how to draw this line would seem to determine who we count as like us, whoever we are, and thus it might describe our moral categories, who is good or bad; Or where and how we draw this line might explain or describe how we interpret the world and ourselves, determining our social orders or psychological identity. This means that when I ask myself what counts as human I begin circling around a (the) confusion between citizenship and personhood. This confusion is acknowledged in our use of the concept of rights. If I recognize someone's rights (as a human being), do I recognize someone as human? Didn't I recognize her human form, or did I imagine she was a machine? My acknowledgmentmeansthatIincludeherwithinmycommunity. Thisinclusioniswhat wecallcitizenship. Myacknowledgmentofsomeoneelseashumanonlyservestoinclude
that person as an example of my kind of humanness determined by his or her inclusion within the community I recognize myselfas functioning in. Thus, the American Bill of Rights acknowledges that rights must be given, and, therefore, that they are not human rights,butrightsofcitizenship. Andtheapplicationoftheserightsisafunctionof interpretation.
The confusion between citizenship and personhood might be necessary for any democracy. Thisconfusion,however,canonlybeansweredbyinterpretations, interpretations about origins, identity, values, marks, etc. ; but the fact that these categories are slippery makes any interpretation (positive or negative) unstable, and thus induces further anxiety about what constitutes humanness.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
49
? As human beings, our questions or motives in determining the meaning of a thing arise because o f the claim things have on us, a claim that describes the material world as somekindoflimit. Thisclaimisnothingnew. FrankensteinhaslittletodowithAIor science, except as an expression of anxiety about our interpretations of the instability of the distinction between inanimate and animate, between human and otherness. It is True that science explicitly questions this distinction, claiming that it is not given but is made, a product of a set of causes and effects that we now call evolution. But the book is not questioning these limits in relation to how they are constituted within nature (or out). Rather it examines the moral implications o f drawing the difference between human and others superficially, examining how we interpret our obligations to others. The problem o f alienation Frankenstein explores is possible because we can identify with the monster and the monster with us: and in this it is a novel asking about the relationship between
personhood and citizenship. It may suggest that our recognition as human (our inclusion in a community) must arise from a recognition of certain origins, but we are not made like the monster and yet he may describe our condition. The novel does not ask but leaves us with the question 'what does it mean to be made'? That question was partially answered and given a special claim on us by Darwin. But how would Darwin fit into a discussion about what counts as human? . His thinking explains, and is now used to explain, how things become living, how living creatures become intelligent, even become human. Do we imagine Darwin would pressure our texts away from the questions o f interpretation exemplified in the confusion between citizenship and personhood? I have to leave this question hanging, but the difficulty of fitting Darwinian thought (as opposed to rhetoric) is
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
? exactly the difficulty o f making the relation and difference between cause (what science describes) and interpretation (what literature demands) clear. This is the difficulty of constructing an intelligence in a machine.
For example how would our claims about the limits or status of subjectivity or our ideological judgments survive philosopher Daniel Dennett's description o f the necessity for selfishness within evolution? :
"As soon as something gets into the business of self-preservation, boundaries become important, for ifyou are setting out to preserve yourself you don't want to squander effort trying to preserve the whole world: you draw the line" {Consciousness Explained, 174)
The interpretations that function within, as and between the confusion between citizenship and personhood in Frankenstein mean we ask for the distinction between the animate and the inanimate, as it were, from inside the security o f acting as animate creatures.
AI turns this interpretive investigation o f the inanimate around. It investigates the difference between the inanimate and animate (or conscious and mechanical) as it were from the outside, not from the assumption that this distinction is meaningful but that it is caused: human beings or animals or plants are effects. Cognitive Science asks 'how can onegeneratequalitativeexperiencefromwhatcanonlybedescribedquantitatively? Itis, for example, much easier to teach a computer how to do calculus, than it is to teach it what a chair is. How can we get from our interpretations of ourselves to the causal mechanisms of our minds? This is, however, as much a question about and of poetry as about AI.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
51
? Keats' ends the "Ode on a Grecian Urn" with the equation: "'Beauty is truth, truthbeauty. '" Evenifonethinkthisisafailedline,theinterrogationofthemeaningofa thing, o f the urn in Keats' ode, suggests that any other answer would fail to give the thing meaning. Even this answer gives the proper form without the necessary force that would make it true (although it might seem beautiful or it might not). But that this question, 'what is the meaning o f a thing? ', should be important to us organizes one o f the currents in the following pages, and at the very least describes a problem romantic poetry found itselffacedwith. Theequationbeauty=truthdescribes,Ithink,arequirementfororan ideal (of whatever status) of poetry in general, although how to construe truth and beauty here is always at issue. Poetry always seems confined by an asymmetry: it can enact meaning or beauty but it's claims on truth are tenuous, becoming more tenuous within modem scientific culture. Things or objects can function as the grounds for truth, and one can argue that this is always the case, but it is certainly the case when what is real is primarily described as material. Within the Keats' ode, the status of the um is complex and shifting (within the poets interrogation), but it functions at least as the vanishing point ofthereal. Thus,theasymmetryofpoetryispartlycounteredbythefactthatitistheum that speaks the equation between beauty and truth: in a rough allegory the thing, the um, with its status as true (or real) asserts an equation between beauty and truth that can only function aesthetically, as an expression of beauty. The um as thing (the real) articulates the truth as an aesthetic. The attempt to enact the qualitative (meaning, beauty, intention, and so on, what AI calls qualia, what it feels like to be conscious or to be human) through an embodiment in language describes what Coleridge called a philosophic poem: to enact
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
52
? the beautiful (or meaningful or intentional) as real and true: poetry with ontological force, the force ofthe real which we describe as true. This is close to what is attempted in AI: to enact the equation between beauty (or the meaningful) and truth, and in this translate our interpretative relations with causal processes.
Traditional AI makes two founding assumptions about this project: what is called the symbolic system hypothesis and the algorithmic assumption. The first claims that the processes underlying intelligence are not just described symbolically but that they are symbolic: we ourselves as well as our thinking are constituted as nothing more than the self-reflexive (to varying degrees) processing of different kinds of symbols within a vast system of symbolic relations. The second claims that our mental states are algorithms, procedures that convert inputs into outputs according to a defined set of rules: a recipe or a function converting or computing inputs into outputs.
What kind of claim do these algorithms and the symbolic processes they enact have on us? AI assumes they are both true and are the only way anything can be made meaningful. In this picture, only symbolic processes, which are not interpretations, allegories or non-symbolic computations, can have sense, and thus our brains must enact these symbolic processes as our mind (There are ways even of construing all of reality as such symbolic processes within physics).
Heidegger wants to displace the claim of this picture of the world on us. Although I do not think Heidegger understood it in this way, the best sense I can make of his claim, in "Das Ding", that "the atom bomb and its explosion are the mere final emission of what has long since taken place, has already happened" ("The Thing", 166) is that the
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
53
? possibilities determined by thinking, knowledge, technology, and economics that produced a working bomb alters the kind of metaphysical possibilities in and through which we function. We can literally enact our death through our knowledge and its potential to destroy the world: I call this a metaphysical fact because our knowledge and knowing, when instituted within social power, has ontological force so great as to destroy us. The worldcanbecontainedbyourknowledgebecauseitcanbedestroyedbyit. Theatomic bomb as a metaphysical fact is a new conceptual limit determining the scope o f our commitments toward what we count as real. We can only believe in ghosts to our greater peril. Thisfact,describingthelimitoftheworldbyourknowledge,enactedthroughthe possibility of nuclear destruction, is that what counts as matter, and what matters and counts for us will have to include a commitment to a physics that can manipulate the substance o f the universe through a complex set o f mathematical models. This is called the possibility o f losing my life by losing the world. That this loss o f the world can be enacted
through human knowing, as if from a single source, gives knowing ontological force not unlike turning the world into a fantasy. What Joyce described as "lethelulled between explosion and reexplosion . . . from grosskopp to megapod, embalmed, o f grand age, rich indeathanticipated"(FW78. 04-06)hasgainedanewallegoricaltarget: wecanliterally enact our death through our knowledge o f the world as a whole and its potential to destroy the world. The problem ofthe machine is not only about our knowledge of how to be mistaken for a thing or how to be mistaken for being human. Our understanding of ourselves and the world must be a kind of reverse evolution, undoing our feelings and interpretations into quantities within some created conceptual language, a language not
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
54
? restricted by the limits o f cause or interpretation. The failure to find such a language has been the failure o f AI.
2. 3 Spiraling from interpretation to cause
The fundamental assumption for Cognitive Science is that the brain is a computer. All forms o f meaning and understanding must be derived from the computations o f neurons. 7 This Computational model grounds all mental functioning in rationality. But this rationality should not be mistaken for what we often oppose by the word
'irrationality' (these words are used primarily as interpretations). Cognitive Philosopher John Pollock defines rationality as the ordered, non-random transition between mental states: "try to imagine a creature possessing mental states but entirely unconstrained by considerationsofrationality"(HowtoBuildaPerson,70) Suchrandomtransitionswould resultinasystemscrashonanycognitivelevelonwhichtheyoccurred. Thisuseof rationality is really a way of arguing that all mental processes (metal states and their
transitions), including language instantiate an ordered system.
Pollock is attempting to construct this kind of system. He tells a story about how
to move from things and causes into interpretations and qualities; this describes an aesthetic,butnotanaestheticlikethatinanovel. ThisisthestoryofOscar. (Iwillgive only the schematics o f this story without concerning myself here with the justification o f the type/token mentalism for which he argues, although such a theory can also be understood to form an aesthetic)
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
55
? Oscar I is a system that consists o f sensory input, some information processing mechanism; which produces a picture ofthe world. This picture is generated from generalizations (computational processes) about this sensory input. These cognitive abilities are embedded in a conative (from Latin conatus, endeavor, attempt) function determining a set o f goals. Such a mental model, although Pollock ignores this, already figures symbolically a kind o f self but in this case one determined only by its pain sensors and its response. If pain then flight, if flight fails, then fight. This describes the mental functioning o f an amoebae, and therefore has its limits.
Oscar H
TheprimaryfailureofOscarIisitsvictimizationbyitsenvironment. Survival
potential increases with the ability to make predictions about the world (or really about itself about its responses to its sensory inputs). Such predictions mean that Oscar II, in replacing Oscar I, must be able to make generalizations about the environment in relation to his goals (in this case only avoid pain). This is the beginning o f mental fiction.
These predictions are generated through self-reflection: an internal sensor senses whenandwhy:pointythingsmeanpain. Isay'meanpain'becauseitoperatesasa
judgment based on a correlation between a rudimentary self (that defined by pain sensors) and the representation of pointy things. This correlation is possible by virtue of the goal (avoid pain) and the possibility for Oscar o f modeling his own sensory outputs. What these sensors do is construct generalizations about certain aspects of the environment in relationtoOscarsownresponses(pain). Suchgeneralizations,reallypossibilities,are
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
56
? logical fictions. There is no question whether they are true or false: everything that looks pointycausespain. porOscarIItheworldisfullofpaincausingthingstowhichhe always reacts as if the possibilities, the generalizations about how the world effects Oscars pain sensors, are always true (but in fact Oscar has no notion o f truth because there exists nopossibility,forhim,ofthesepredictionsbeingfalse,eveniftheyarefalse). Possibility means always true fiction. This is the mental world o f a worm.
Oscar IU
The problem with being a worm is that the environment can be deceptive and one's own sensors can fail or produce false information. Not all pointy things are alike, a pointy leaf is not like a pointy stick, and a machine-eating tiger is not the same as the mirror image of such a tiger.
Oscar HI experiences a mirror stage. Initially Oscar I and H were like amoebae to worm; Oscar n and m , however, are like bird to cat (I guess the bird has eaten the worm). The cat is able to distinguish a mirror image from another animal (it's not clear whether the cat thinks it's its own image). The bird responds to a mirror image as if it is another bird (it attacks the image)
Oscar HI must be able to evaluate the ontological status of its perceptions, are they real or not, it must be able to recognize mirrors as mirrors, and mirror-images as mirror- images. Oscar n sees the world phenomenological, as if its eyes were video or film cameras. If you hold a camera and film while walking around the picture jumps, and bounces and shakes. The viewer gets a headache trying to stabilize the picture into sense.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
57
? Why don't we see like that? We walk around and yet our picture of the world doesn't really jump around like this; why?
To answer this we must ask how can we make Oscar m generate a world separate from his senses, such that he can evaluate his sensory inputs in relation to this stable world? Pollock ask: how can Oscar distinguish between machine-eating tigers and mirror-image tigers? Asking this question means asking the difference between appearance and reality, and its answer requires an interpretation.
This level o f interpretation requires another layer o f internal sensors, this time evaluating the first level o f internal sensors. Pollock thinks that our picture o f the world is generated by two different kinds ofgeneralizations: "On the one hand Oscar HI has generalizations about the relations between his perceptual inputs and the states o f his environment, and on the other he had has generalizations about regularities within the environment that persist independently of perception of the environment"(4). This independent world is generated as a set of generalizations not about the world but about Oscars correlation o f himself (defined by avoidance o f pain) to representations (processed sensoryinput). Thiscorrelationisevaluatedinordertodeterminehigh-levelregularities that constitute a stable world, that in essence allows us to see without all the bouncing.
Now these kinds of evaluations take place at many different computational levels within the brain, and most o f these evaluations take place through hard-wired computational processes about which we have no awareness.
Pollock argues that there is a distinction between sensation and the feel of sensation. Thefeelofsensationistheoutputofsecondlevelinternalsensors,sensorsthat
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
58
? sense the first level internal sensors that process our perceptual information: being-aware means and is the sensing o f perceptual input (this is the link AI wants to make between 'mean'and'is'). Thequalitativefeelofourconsciousnessorsensingmeansandisthe processing or sensing of lower level internal sensors by higher level sensors. Consciousness and the qualitative feel o f our experience, what understanding feels like, or meaning, or love, or whatever are only what we call this second level sensory processing of our own internal processing.
I think this implies for example that all o f our emotions, love, hate, desire, fear, etc. are only interpretations of physical states. We try to make meaningful (sometimes by connecting it to some putative cause) physical states that we sense but do not understand. Even our emotions are interpretive fictions, convenient descriptions that may mask a mechanism whose primary purpose may be to simplify decision processes (emotions force us into decisions). *
What exactly does it mean to process the information from our first level internal sensor? Pollockshiftstheproblemofconsciousnessorqualitativefeelingintosome magical processing. How do these interpretations produce new effects? Or how do we invest ourselves in all o f the complicated ways we do in our correlation between internal sensors and our picture ofthe world? I don't think we should find Pollock's picture satisfactory.
Whatever the merits ofthis model ofthe mind (and there are many), it leaves us with a puzzle. So far I have described only two kinds of meta-languages, which seem to describe the limits o f the other, but whose relation remains unclear: a language o f cause
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
59
? described vaguely here by different versions of Oscar, call this Darwinian engineering, and a language o f interpretation, commonly used in literary analysis, asking about meaning, anddeterminingthedistinctionsbetweenanimateandinanimate,etc. Butthisinterpretive language undermines its own status by showing how its distinctions and categories are so unstable as to be senseless without ontological limits o f exactly the sort described by science. But these scientific limits are not foundations precisely because we cannot get outside o f our interpretations. Thus these languages o f science and literature speak in two directionsatthesamething. Ithinkthisisourinheritanceandtheforgingofanother language that includes both of these languages is the underlying although unrealized promise o f AI, and I think our problem.
1Translated and found in Kenny, "Wittgenstein on the Nature of Philosophy", 14.
2 In Being and Time.
3 Is Thoreau's jump from anyone's embodied, particular eyes to God's eyes justified by the same trick that Heidegger uses to move from Dasein-as-mine to its identity as Being-with in Chapter IV of Being and Time! The kind of holistic, non-formalizable link between particular and universal that Heidegger constructs both elucidates a critical dependence between meaning and existence, while at the same time offeringthegreatestresistancetothepossibilityofembodyingaconstructedmind. Daseinbecomesa soul. Ibelieve,however,thatHeidegger's"soul''isfalse. (Howisitpossibletotalkaboutfalsesouls? ) But I am nevertheless left with investigating the intransigent existence and meaning of our use of, belief in,andneedforwhatHeideggerpointsto. Ifconceptsarenotthings,thenwhatarethey: howisit possible for us to function at a level o f abstraction beyond materiality?
4 Cora Diamond describes what it means to understand the answer to a Sphiax-like riddle:
[T]o know the solution to a riddle is not merely to know o f something that thought o f in some way or other it is the solution: you have to know how it is the solution, If you are simply told that man is the solution to the Sphinx's riddle, but do not understand why--ifi that is, you do not understnad why man is supposed to have four legs in the morning and so on--you are in the positoion of someone who 'knows' that every equation has a solution without any idea o fhow that may be arrived at (270)
5This is, of course, not the end of the opportunities for analysis.
6 P. M. S. Hacker. Insight and Illusion: Themes in the Philosophy o f Wittgenstein. 1Neuronscanbeintwostates:excitatoryorinhibatory;onoroff:thustheyaredigitalcircuits, by stimulation a neuron reaches an activation potential and fires. The energy of this firing is not variable-- so no information can be encoded in the strength of the output--only by the fact of the firing and in, what doeschange,therateoffiring. Greaterstimulationofaneuronincreasestherateoffiring. Thereissome confusionwhetherthebraindescribedinthiswayislikeadigitalorananologcomputer, digitalswitches (or in computer science these are called logic gates) are like digital clocks, they are in one state or
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
60
? another- analog switches are like a clock with two hands--the changes in rate describe an analog relation. Itdoesn'tmatterforus--buttheassumptionbymostAItheoristsandcognitivescientistsisthat it is digital. .
8 This might be a way o f explaining how replicants in the movie Blade Runner can generate emotions even though they are designed not to have any.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
61
? 3
Romantic Fragments and Modernist Machines
InalettertohissisterandbrotherKeatsimaginessoulsasmadenotgiven. This making proceeds through what is given: the world as circumstance and the mind as intelligenceandmemoryandemotion. Whywouldapictureofthemindexpressa theology or a moral stance? "The point at which Man may arrive is as far as the parallel state in inanimate nature and not further" (326). 1The inanimate is the limit o f human happiness. Thislimitengenderstheworldasa"valeofSoul-making",rewritingthe Biblical "a vale o f tears. " In support for this inanimate limit, Keats animates a rose with sensation, in order to tell a tale of crucifixion:
For instance suppose a rose to have sensation, it blooms on a beautiful
morning, it enjoys itself, but then comes a cold wind, a hot sun--it cannot escape it, it cannot destroy its annoyances--they are as native to the world as itself--no more can man be happy in spite, the worldly elements will prey upon his
nature. (326)
The unity between the divine and human, between fundamental ontological difference, is twisted toward biology, into a unity between rose and human. In order for Keats to tell the story of our inhabitation and involvement in the world as a limitation on our happiness he endows a rose with a conscious o f that which threatens not its consciousness specifically, butitslife. Sucharoseisnotnativetoanyrealworld,butevenitsfantasticsentience
Notes for this chapter are on page 102
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
62
? does not free it from its containment and belonging to a world described by power, not determined by the roses intentions.
The rose slips from the perception ofsensation to an interpretation ofthe meaning ofthese sensations as 'annoyances'. Even to evaluate negative stimuli as annoyances implies a complex o f responses that involves the possibilities o f happiness and hope, a language ofintention in which the consciousness or the desires or beliefs or values ofthe rosestructureitsrelationtotheworld. Therosefunctionsasarosewithinasystemof opposedintentions. Intention,inthiscaseannoyanceorhappiness,isafunctionofthe systematic relation between the rose and the world, and is therefore in neither. The world is animated and the intentionality o f our involvement in the world is determined and guaranteedbythisanimation. Coleridgepicturesthisin"TheEolianHarp":
And what if all of animated nature
Be but organic Harps diversely framed,
That tremble into thought, as o'er them sweeps Plastic and vast, one intellectual breeze,
At once the Soul of each, and God of all?
Thought is generated as a consequence of our involvement with dominant non-human forces;intentionfunctionswithinthesystemofworldandbeings. Ourintentioncanbe generated by reducing ourselves to things--and then displacing intention into the environmentthroughwhichthisintentionbecomesvisible. Weenterintothedominionof arbitrary forces to the degree that we can let the rose stand for ourselves, allegorize ourselvesasusedbytheworld. Inthissensewearealwaysinvisibleinthisallegory. We
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
63
? exist as nothing more than our own interpretations o f ourselves through our language or actions. Allegories describe and constitute a limit to our knowing ourselves.
Consequently, it is a mistake to imagine that these allegories can describe our becominganything. Theycanonlydescribehowourinterpretationschange. Whenthey describe change they describe it teleologically, as if the allegory were a machine. Even if wearemachineswefunctiononlypartiallythroughourinterpretations. 'Causation'hasa particular grammar that overlaps but is not continuous with 'interpretation'. Wittgenstein at one point investigates how the word 'reading' describes how we read words into speech, how words 'cause' our reading, not our understanding o f what we read.
But why do you say that we felt a causal connection? Causation is surely something established by experiments, by observing a regular concomitance of eventsforexample. SohowcouldIsaythatIfeltsomethingwhichisestablished by experiment?
