One aspect of the study of natural language is the quest for a universal grammar, a system that would explain conclusively the way all languages are organized and function. This is in part a search for an underlying principle of organization. This paper contends that a universal grammar is not to be found by the usual method of studying words, abstracting categories of words, and overlaying a structure based on those categories. We should discard this approach altogether and adopt a different technique that emphasizes the biology of language: language as a product of the brain. Grammar, and generally the organizing principles of language, can be seen as sharing biological rules of constraint and are subject to evolutionary principles. Further, language may itself be an expression of similar rules that inform the material processes of evolution; one thus may be able to understand information transfer in living organisms (genotype to phenotype, synaptic chemical communication) by an analysis of natural language.
That language functions reside in the brain is a truism. Early studies of brain damaged patients led to the belief in isolation of brain areas for specific linguistic functions (Broca, Warnke), an approach popularized by Wilder Penfield's experiments of exciting brain areas of patients who would then recall often vividly a particular past event. This gave rise to theories which posited that memories resided in toto in brain cells, the brain a kind of quiescent repository of discrete events neatly compartmentalized. Yet recent technologies such as positron emission tomography have distinguished only very general language areas (Damasio and Damasio), and a more broadly based interaction of different brain domains in response to verbal stimuli.
Penfield's patients, it is now believed, actually confabulated their earlier 'memories,' creating stories engendered by a cascading of associations brought on by the electrode in their brain. All attempts to discover a one-to-one correspondence of brain cells with specific memories have failed, and associated theories of a strict hierarchy of mental functionality have similarly faded. Similarly, the proposition of sharply delineated properties between the left and right brain, while seemingly correctly positing some differentiation between the hemispheres, is now viewed as overly schematized and, like strict localization hypotheses, belies a great diffusiveness of brain activity.
Arguments for more diffuse brain functionality span philosophy, cognitive science, and molecular biology. Encapsulating these is something like Dennet's remarks of consciousness itself: the clear-cut, Aristotelian, Cartesian boundaries of mental categories and mental processes seem incapable of representing the complex, overlapping functions of the brain, an approach which finds support in Darwin as well as writers such as Dawkins and Gardner. "The absolutist or essentialist philosopher is attracted to sharp lines, thresholds, "essences, " and "criteria".... Opposed to this way of thinking is the sort of anti-essentialism that is comfortable with penumbral cases and the lack of strict dividing lines. Since selves and minds and even consciousness itself are biological products...we should expect that the transitions between them and the phenomena that are not them should be gradual, contentious, gerrymandered" ( Dennet, 421; also Rosenfield).
This kind of "gerrymandering" also has been articulated as fundamental to general category formation in human cognition (Lakoff, Johnson; Rosch). Within this emerging model of consciousnessis the recognition that what the brain has been doing affects what it is about to do; that the past and present are intimately bound together; that the brain's functions are largely determined by its structure; and that abstraction is a continuing process within the neural structure. In essence many of the previous top-down theories describing the 'mind' as analogous to neatly ordered building blocks that interlock into a fine edifice are being replaced by more bottom-up descriptions that focus on the neurological properties of the brain. The latter, unlike the solid categorical blocks, view the brain in terms more analogous to interacting gases.
Stein and Meredith have taken this even further, arguing for a prevalence of multisensory integration, and exploding the modality-specific paradigms. Cross-modality appears key to perception in more sophisticated animals (1993). Their theory also posits that this integration takes place at the single neuron level. Growing interest in brain behavior such as synesthesia also suggests that the older models of isolated modal functionality does not adequately explain the brain. Seemingly bizarre examples of people who can taste shapes are not to be viewed as aberrations, but as evidencing a widely shared property or potential that is masked through acculturation (Cytowic).
When seen in this light, traditional grammar, which attempts to discover language rules, has often drawn us away from the biology of the phenomenon and into a metaphysical realm that schematizes and makes brittle a far more volatile and less predictable system of communication. 'Grammar' is usually defined as an aspect of syntactic studies, a generally accepted part of linguistic theory in which syntax and semantics are the chief components. Syntax and semantics are often purposefully kept apart as separate fields of inquiry. The difficulties of this methodology have been contentiously described (such as Givon, 1979). More recently Rosenfield, again echoing the tenets of contemporary consciousness theory, goes as far as to declare, "grammar is an incidental consequence of [the brain's processes of] abstraction. Like subjectivity and like consciousness language thus develops. To view it as a set of abstract grammatical rules falsifies its very nature; it is part of the very structure of consciousness" (112).
The imposition of grammar upon language is generally speaking a top-down enterprise that labels linguistic phenomena in discreet units with rules extrapolated from the relationship of those units. Hence the preponderance of sentence level analysis, usually simple sentences, and an avoidance of poetic diction and metaphor. Chomsky, by point of distinction, looked for deeper, neurological explanations for language rules, but also disconnected grammar from memory (1965).
Rather than pressing for one theory of consciousness over another, or arguing for a particular definition of grammar, for the purposes of this article it is instructive to put aside many inherited assumptions. It is more efficacious simply to assume that language is a part and product of the most complex biological organ; that this organ does not appear reducible to a Cartesian model of a central controller activating and guiding brain productivity in a logical sequence; and that the categories we formulate using language are similarly fluid and not reducible to Aristotelian, hard and fast constructs.
For the purposes of this article it is also instructive to put aside the notion of mind, and rather see 'mind' as an ideological construct that does not exist separately from brain. Similarly, we may discard the concept of grammar as it has come down to us as an ideological construct. This is, in part, a reaction to the general tendency to apply rules to a very small, artificial set (sentence) of discrete words, a process that assumes that a strong discontinuity between words exists and that meaning can be extracted from analyzing patterns that are dependent on preconceived demarcations of word types. Too much research exists to accept strong discontinuities in the brain. If language functions, as I believe it does, by nature of the structure of the brain, then the boundaries of its constituent parts are less predictable than previously imagined, and our edifice of grammar is more of an a priori scaffolding that hides a remarkable and utterly dynamic tool.
In the place of these assumptions, new ones may at least open the way for a more cross- disciplinary approach to language that emphasizes the physicality of the brain and the material basis for language functions. These would include the recognition that the complex functions of metaphor and other sophisticated rhetorical devices are only explained by a brain that 'gerrymanders' in the basicprocesses of cognition; that the structure of language owns analogies to other information exchange mechanisms, such as genetic interpretation when reproducing; that language is a natural expression of evolution that performs an integration of the sensori-motor, emotional (limbic) , and cortical areas of the brain. Language might be seen as an emergent property of life itself, its origin random and its future without inevitable goals.
If we define consciousness as access to the neurology of the brain, and awareness as recognition of this access, then language is the brain utilizing the brain, reinforcing this access by its appropriation of it. Language 'stains' brain function while reconstructing it; biological structure and the syntax and semantics of language are inextricably twined.
Some brief extensions of this line of thinking follow.
The Neurobiology of Language
Language resides on the neurological arrangement of the brain and this neural wiring is somehow utilized in the process of language generation and reception. Using the lexicon of molecular biology, one might posit that instead of nouns and verbs and objects, language generation entails in part the strength of ion transference across synaptic clefts. To view a schematic of a neuron, with its dendrites and synapses, is to 'see' grammar and, should we capture the actual exchange of molecules across the cleft, to 'see' language at work. This assumption poses immediate difficulties, as there exist approximately 100 billion neurons in the brain, each of which is extraordinarily complex in its composition. Compounding this, we also know that neurons are connected to other neurons in widely different patterns which can be determined by an individual's experience, and that to trace a cause or effect, or to trace the connectivity of a neuron, is a labyrinthine affair. As noted above, strict localization of information or processes seems doubtful.
There are also different types of synaptic transmission (voltage-gated, ligand-gated); many kinds of neurotransmitters (glutamate, the catecholamines, and serotonin, for example). There also exists a system of neuropeptides that play different roles in brain information processing (cholecystokinin) and more simplified chemicals (acetylcholine). The act of transmittal from synaptic terminal to receiver frequently involves additional stages of biosynthesis. "Second messengers" may form, adding a further modulation onto the process.
An elementary schematic of neurotransmission inspires a few observations. Firstly, it is obvious that even with the technical tools of the late twentieth century, the material brain remains hauntingly complex. Indeed, because of the sophistication of modern analysis the complexity has become better known and appreciated. Secondly, when discussing language, we may also begin to appreciate why the most complex human tool ever devised is so complex. This in itself argues for a 'bottom up' theory of language generation and interpretation. Language has rules, but these rules are a) themselves made up of the material brain, i.e., are structuring devices that ultimately reside in the learned correspondence of a vast array of neurons b) operate along the material neuronal pathways of the brain and c) are thus determined by the biological structure of the brain.
Acknowledging this, it is much easier to understand how language can alter moods, change attitudes, influence behavioral patterns, or result in deep conceptual conversion. Words can cause anger, peacefulness, elevate pulse frequency, precipitate the hot rush of adrenaline surges or make us feel colder by inducing fright. Language can act as a chemical change because it is itself comprised of chemical changes. Biosynthesis is required at some level to read this page, just as it was required to compose it. Because of the strong moods and emotions language can induce, it is probable that language uses both the cortical and limbic system in the brain (Cytowic, 1989). Words are not incorporeal.
What does this mean for the understanding of language? The brain itself is the product of evolutionary biology; each of our brains has developed from birth in what can be described as a highly complex sequencing of information transfer that begins with DNA protein and advances throughvarious interpretation and translation procedures to build the brain as well as other differentiated organs. In many accounts of human development at the molecular level, terms usually associated with language study are used, such as translation, transcription, information exchange, and semantics.
The creation and growth of a biological life form is a series of re-interpretations of code that can result in differentiation of parts that function coherently. We speak of the 'language of genes,' and the 'language of proteins,' but much more rarely of the protein components, or the salts, or synaptic gates of War and Peace. Rather than impose upon language a further abstraction, we might see language as a result of these biological processes that should reflect in some ways these processes. The rules of language would not thus be innate linguistic rules for language generation that are separate from other kinds of brain function, but generalities and predispositions that govern evolution and other instances of information exchange in the living organism.
One correlative, however, we do ascribe to information transfer whether it is via proteins or literary texts is the use of symbols. Natural language is symbolic, as it is representative of something it describes, and strings of natural language components--the context of word groupings--can further symbolize. Thus the word 'chat' in Madame Bovary may symbolize on one level a living feline, or a category of felines, and with repeated contextual occurrences in the novel also come to symbolize infidelity. That the relation between the genotype and phenotype is largely symbolic may also have bearing here, just as the observation that only matter-symbol entities evolve.
Evolution, like natural language, is continually adapting. Pattee, for instance, has long noted these similarities, especially in light of his semantic closure hypothesis, finding that "the multiplicity of function in evolution as well as the multiplicity of meaning in natural language support...a multiple- closure concept of representation and meaning" (1982, 339). This underscores the highly flexible nature of information transfer in biological systems, stresses the possibility of meanings rather than a fixed or limited set of meanings, and draws attention to the conceptual affinities of language and the biological processes that have ultimately produced it. The use of symbols to produce meaning, and to reproduce biologically, in fact seems inviolate. A symbol both reproduces itself and requires a translation by linking to and referring to something else at the same time. The symbols of language occupy matter, and use matter for expression. Life itself, viewed as such, may be defined by the ability to transform matter into symbols.
Properties of Words
Because it is impossible to do justice to the nature of language in such short space, focusing on the aspect noted above of joining together different material parts for the expression of meaning must suffice for the purposes of this article. Even a brief and representative description of the bewildering potential of associations words allow underscores the intricacies of meaning in a word or series of words. It is important to note that these associations are hardly limited to other words, but entail a plethora of ideational, tonal, rhythmic, emotional, and intentional contexts.
Ancient civilizations identified numerous phenomena by which words associate, and thereby become redefined, either through sound similarities, metrical rhythms, or conceptual proclivities. Many myths are elaborate examples of new conceptual associations being drawn. Counter to many popularized versions of myths, they more often portray an unsettling metamorphosis in creating new representations and symbols to explain the human state. Less elaborate examples are rhetorical devices: homonyms (hear/here), assonance (open the door slowly), alliteration (the forest's ferny floor), rhyme itself (Stein/ rein), or simple repetition. The richness and sophistication of word 'overlapping,' the myriad associations words can have with one another, and the fact that intricate concepts and categories can be triggered by single words (love, death), appear naturally to correlate with our evolved biological system.
Certain etymologies are similarly intriguing. They display a somewhat different type of association-- ideational development and semantic nuance-- over time. For instance, in English, the words conscious, science, schism, and schizo have the same Indo-European (I.E.) root, *skei, whichmean to cut or split. 'Schizo' and 'schism' seem more obvious; science and conscious most likely derive from the act of separating components and making distinctions. Splitting is requisite also for category formation and new knowledge acquisition. 'Burden,' 'birth,' 'confer,' and 'differ' also share an I.E. root, *bher, meaning to bear. This came to be associated, in a visually understandable way, with carrying. 'Confer' is synonymous with 'holding a conversation,' while 'differ' means 'holding apart.' The I.E. *ar- , meaning fit together, is also interesting, if less agreed upon. It has come down in English as arm (the weapon) and armor, which are objects fitted together. Some also believe this root to be related to re- and ri-, whence riddle and even a verb meaning to think, both actions involving, metaphorically, a fusion of disparate elements into a recognizable new whole or category.
Equally sophisticated are rhetorical devices such as synecdochy, metonymy, simile and, perhaps the most studied, metaphor. Each brings a revelation about the capacity to extract and map multifaceted linkages as understandable associations, and thereby nuances and meaning shifts, in almost any word or word construction. With synecdoche a part can substitute for a whole, as in "all hands on deck" (the entire person is extrapolated from this phrase, but the choice of 'hands' suggests the key body part in the performance of a sailor's duties). In metalepsis a metaphor is substituted for a cause: "he cannot speak the truth; curse the wood from which his cradle came." The association is 'distant' and not readily obvious, but enough semantic overlap exists wherein we might interpret wood and cradle as symbolic of this liar's upbringing or environment.
Metaphor offers unique and, of late, widely researched suggestions about the way language and cognition operate. From the Greek for 'transference,' it is generally agreed that a metaphor brings together two or more similar yet dissimilar entities and yokes them together in unexpected ways. It is important to note that metaphors engender emotional response (Mac Cormack), for this is part of their power. They are also tools for discovery of new knowledge and rethinking the way we understand (Rifkin, White). Aristotle recognized metaphor as a means to discover new knowledge. It is perhaps surprising to learn that not very long ago metaphors were considered an aberration or dysfunction of language.
The astounding ability of metaphors in regard to intellectual stimulation and the bridgement to new understanding is perhaps best appreciated in poetry. For example, a single phrase from W.B. Yeats's poem "Byzantium:" "the dolphin-torn, the gong-tormented sea". Technically it is not a sentence, but it has regularity of structure because of the verbal stimulus, that is, what the five words connect. It is metaphor-like, in that we can extrapolate "the sea that is dolphin-torn and gong- tormented." The phrase is utterly untrue in some respect: as a liquid the sea is not torn, as an inanimate object it cannot suffer torment.
But it is also precisely true: the sea can be described as passive and inanimate yet is continually beset by leaping life and time itself, or ritual, encoded in the sound of the gong. Within five words are conjured life and the means we have to transcend it, with the tacit assumption that we want to transcend it because we eventually die. We as human beings are similarly beset, as much as some of us wish for the timeless, mythological existence of an untroubled sea, which is only suggested by the phrase.
Further extrapolating , we can also appreciate that a metaphor and other figurative language works because of the following capabilities: it resonates between the literal and the non-literal. A tension exists that is an emotional response to a perplexing juxtaposition that relies on similarities for its understanding and dissimilarities for its meaning. In this respect a metaphor is a composite of available and currently utilized information and that which is new. As conflicts collide there is transference (a new understanding is achieved) and thereby transcendence, a key aspect of metaphor ( Harries).
Schematically, the following are attributes of metaphor:
a. overlapping of one or more categories associated with words, and clusters of words
b. the production thereby of a new category
c. tension between concepts entailed in the overlapping categories
d. missing information (the correlations are never exact or clear-cut)
e. self-similarities are established
f. transcendence occurs; semantics must entail transcendence
g. boundaries of meaning are fluid
h. issues of truth and falsity are blurred; a metaphor creates an included middle that is an amalgam of its constituent components
i. non-literal aspects (ideational, emotional, personal history) are important
in the understanding of metaphor
j. transcription between one object and another takes place
Here is evidence of a system that picks and pulls from a variety of shared affinities--sound, ideational component, and more sophisticated cognitive groupings--in its production and interpretation. It yokes together elements, often illogically or counter intuitively, that by virtue of their joining can represent one another, further extending the meaning potential of individual sounds or words or phrases or sentences. Rather than seeing metaphor as an aberrant aspect of language, it is more productive to see it as an intensive example of properties common to language in general. Sentences are not lists; words are not stored as in a dictionary or thesaurus. Language does not stand still, but resonates continually as meaning is recreated along general but not completely predictable guidelines.
If we view the process of the creation of new biological life through DNA transcription as a form of narrative, then the correspondence between language and general evolutionary processes becomes even more suggestive. In the production of new life a story is enacted; the four letter alphabet of nucleic acids is translated into the twenty four letter system of proteins. Codons form to encode amino acids, and, equally as important, codons form to terminate the sequencing. Without the terminus (closure) there would be no life possible. The substance in DNA becomes information when it is transformed into instructions. It is reassembled into symbols. This symbolic property is emergent: it is not something that can be necessarily intuited, but comes about because of reassembly. Through its symbolic property a phenotype is produced.
Also suggestive of the discussion above, the act of transcription is now known not to be a strictly regulated pairing, but involves instances of what is termed 'RNA editing.' This can entail transforming or adding to different RNA bases. The RNA in mitochondria and chloroplasts is different from that of RNA in the nucleus. The AGA codon in mitochondrial RNA adds valine to a peptide chain. In nuclear RNA it is a stop codon.
In this extraordinary series of processes flexibility, editing (reinterpreting) and, in the example of mitochondrial RNA and that from the nucleus (identical material with entirely different causal effects depending on the context), multiple meanings are in evidence. The act of splitting and transcription gives rise to a slightly altered, or nuanced new sequence of the original DNA. This will combine with similarly nuanced DNA to begin a new life form. A literal transcription or translation does not exist; the absence of point-to-point replicability is most likely critical to evolution just as it is critical to the production of metaphor (Genetica, 86, vol. 1-3).
The new life form, or phenotype, may be described as a metaphor of the genotype. As a union of two disparate but similar phenomena, a metaphor exists, as noted above, as a combination that both partakes of and is distinct from its combinatorial elements. As a means to discover new knowledge and expand upon existing paradigms of understanding, a metaphor is a critical tool. It would not be surprising that it is echoed in the processes that conduce to biological life. More theoretically, the presence of an exact, literal replication (perhaps at the level of bacteria) would entail something akin to eternity, as differentiation is impossible and distinction, with its temporal associations, would be non- existent. Metaphor, and the associated properties of language, thus may be bound inextricably to the notion of mortality and the distinctive editing of new life. Closure may itself entail meaning, or perhaps meaning is an emergent property of closure.
The Fractal Nature of Grammar
Extending the phenomenon of metaphor, it is important to better correlate other properties of language to its biological platform. The above discussion of metaphor and language associations may have suggested a connectionist argument , which was not the intent. Because of the multiplicity of associations contained in a word, and the myriad further associations triggered by coupling words together, any point-to point-connection seems impossibly simplified. Some theorists have attempted to define the more complex kinds of associations inherent in language by terms such as 'phase space' or 'semantic space' that can in turn be stimulated into trajectories that create semantic vectors.
These concepts derive in part from work concerning the sensori-motor area of the brain in response to the difficulty of explaining how the neuronal activity at the cellular level, even when it is very well understood, translates into visible behavior like catching a ball or uttering a sentence: what is the mechanism that coordinates millions of activated brain cells into coherent behavioral patterns? Descriptions of phase space and the translation of vector to vector (e.g., visual to muscular) in this case owe much to the work of Pellionisz and Llinas (1979, 1985; see also P. Churchland, 1986, 412-433; MacCormack).
Such terms can be put aside. It is not necessary to in fact identify where or how a word is posited in the brain, but rather to emphasize the amazing fragmentation involved in producing meaning. The term 'word' may be linked to 'whirred;' it may have other linkages such as 'Ur' or 'worried;' it may conjure mythological or theological associations, and in turn reflect highly ideational components ('In the beginning was the Word'... the beginning of time, of the earth) or images of books, or the concept of authority (the 'final word').
In this respect it is also misleading to talk about the meaning of a word as if it were contained in the word, as if the phase space entailed all these concepts and associations. Rather, meaning is the potential associations of a given word, or phoneme, or combination of words. Understanding the vast amount of previous and new potential associations, a word at any one time can have a variety of potential meanings while having no meaning at all. Meaning in this sense is a neuronal confederacy of influences with inherent flux; it is impossible to simply define a word at any given moment.
When words are combined, what is the coherency of organization? The trajectories of associations are myriad, yet without some agreed upon structure the potential meaning is lost as babble or becomes like the strangely patterned but ultimately incomprehensible expression of madness. Within its coherent organization, the complexity and highly fragmented nature of meaning is evidenced in the phenomena of riddles, jokes, puns, word games, and the hundreds of contending interpretations of Hamlet. This is not to say that every interpretation or meaning has legitimacy or equal weight. We have, as a species, constructed generalized sets of appropriate influences by which the brain is utilized for general communication. Within this generalized set, however, is a nearly boundless potential for nuance and new discovery of associations, which can perplex, shock, confound, or delight us.
Research on the geometry of language is highly suggestive in this regard (Van Fraasen, 1980 Van Fraasen and Hooker, 1976 ). If indeed the brain has learned to utilize itself, and this is what we pass along generation to generation, and why knowledge cannot be inherited. Our children inherit only the structure and with it the potentialities of re-structure.
A phenomenon that occurs in nature and can be modelled mathematically that might shed enormous light on the function of language and, specifically, the organizational property of grammar is fractal in nature. If understood as a fractal, grammar would more easily be explained as an ever changing morphological set of basic instructions. Like semantics, indeed not to be distinguished from semantics, grammatical organization as a fractal phenomenon is a shifting yet stable structuring device that generates word constructions but is also the essence of the meaning. That the neurological arrangement of the brain is also fractal in nature is no coincidence. In this regard it is impossible to separate the dancer from the dance: to see grammar as an abstraction removed from meaning.
Language activates these associations as well as produces them. In one regard the brain, through language, is constantly in search of new meaning. Meaning is thus highly determined by initial conditions (another feature of biological phenomena). This raises the issue of symbols and symbolism: what is the symbolic nature of language; does language, as a system of symbols, manipulate symbols and create new ones? Earlier descriptions of language would have words as symbolic representations of objects external to the body, or symbolic abstractions of qualities such as faith and honor that can become evident through behavior.
It is only in the act of assembly that language achieves the threshold of symbolism. To an English speaker, 'ba' and 'ka' (unlike an ancient Egyptian) means very little. Patients with neurological damage will often say these syllables back as 'bake' or 'cake' (Ornstein) because they, like possessors of undamaged brains, strive to bring linguistic components into a symbolic range. I would posit that, as a product of evolution, we are continually, inexhaustibly, reworking received stimuli into symbolic meaning.
To extend this argument somewhat further: language does not exist, at least as a moment-to- moment phenomenon along a temporal continuum. Particles of linguistic phenomena reside in memory and most likely have a plethora of associations with other linguistic particles, images (or parts of images that can be similarly assembled), moods, smells, and other contextual linkages. Language is made manifest as a combination of its parts. Its potential exists equivalent to the strength of influences in the neurobiological netting of the brain and the capability of new re-formation in a fractal-like organization. As we most likely re-create ourselves moment to moment, awake and in dreams, we reconstruct language.
Fractal Grammar and Symbols
If this argument owns validity, then one can posit that each word has an associated grammar. Each word is thus fractally linked to other concepts as part of its natural biological state. The stringing together of words into phrases or sentences does not produce grammar, nor do sentences or phrases rely on grammar as an external organizational principle for structure. It contains in its parts the fractal property that determines its associations with a great deal of flexibility.
This would suggest that symbols are inherent in human language-- they are inseparable from the form and meaning of language. A symbol by definition throws characteristics together, something like a more extended, structured, yet reduced metaphor. To understand a symbol is thus to understand in part the brain. It is wired and forks, and links in billions of potential confederacies. Without that linkage there is no meaning. Because the neurobiological threads of the brain fork and separate in a fractal way like ever budding trees, the possibility of identical structures seems implausible. Grammar is a time-dependent approximation, a formula that yields different meanings as the context changes. Being time-dependent (past usage, continual modifications) the context has to change. Symbols perform a stability of meaning in this environment, but again, the creation of a symbol is apparently directly the result of the organizational network of the brain.
A future paper will explore the question of grammar itself as functional symbols.
Artificial Intelligence and Artificial Life
The computer modeling of language has long been the domain of AI. Researchers have attempted to simulate the thinking process and to develop programs that function along the lines of natural language. It is interesting that Turing's famous test of an intelligent machine is essentially an exercise in language generation and interpretation. If someone on one side of a curtain cannot tell the difference between the responses of a human and a machine behind the curtain, then the machine is indistinguishable from the human and thus can be said to possess intelligence.
There are a number of critical flaws in this 'test,' but it is at least a quintessential expression of the top-down approach. If language, to be interpreted, requires the assembly of parts that, ofthemselves, have no specific logical relation, into a reformation of matter (cells, air, muscle) that produces symbols as an emergent property of the assemblage along flexible guidelines and rules absent of a blueprint or overarching plan, then it is difficult to see how computers, as they are programmed for AI, can approximate this phenomenon. Turing's machine is, even if we imagine it to perform with the greatest facility, an artificial language expression device, with limited applications because it is based on an a priori 'proof.'
More promising for the understanding of language is the research in artificial life (A-life). Here simple algorithms can produce complex behaviors that manifest in shapes and images on the computer screen. Patterns of development over many generations can appear, parasites can be programmed in, and models of biological evolution can be produced from basically simple initial conditions.
Similarly, language appears to generate from simple rules into immensely complex expressions that involves the transcendence of fixed matter into matter that assumes symbolic properties. In this respect matter represents itself. Existence owns perhaps far more of a symbolic or metaphorical quality than we might otherwise imagine. This is not to argue for 'metaphysical realms,' by any means, though the inclination to suspect that they exist is understandable given the processes of evolutionary development and reproduction.
As others have more recently argued, it is important to break away from Cartesian (and Aristotelian) assumption of categories of order and hierarchy in the brain, as well as to disclaim the Platonic principle of an ideal representation in the 'mind' that the external world only reflects in an impoverished, shadowy way. Plato and Aristotle are much closer in their philosophies than often credited, and both have contributed to an important tradition of misinterpretation of language. Too often Grammar, as a defining structure informing all language, echoes Plato as an idealized mental phenomenon. The particulate approach of word and sound categories and the way they 'stack' upon one another is Aristotelian. Both aspects tend to exclude biological nature from consideration, and language is without doubt a purely biological product.
As Edelman has more recently argued (1992) it is imperative to introduce biology into the study of mind. Following Lakoff, Edelman sees language formation and use as closely aligned to his own theory of neuronal group selection (1985) in that language is dependent on conceptual categories and is not 'generated,' in the Chomskian sense, by a language 'device' in the brain. It is also interesting to note Pollack's work on DNA and its interpretation: his fear is that rigid schematization and narrow focus when documenting the genotype will yield potentially dangerous ideas concerning the concept of being human-- and what genetic engineering might make more 'efficient.' Pollack argues for the richness and diversity of the genotype, DNA as history and susceptible of multiple interpretations like any historical text. Both Edelman and Pollack essentially wish for a more programmatic marriage of traditionally separate scientific and humanistic disciplines in order to enrich both.
Neuronal groupings and DNA as text are important features of a growing awareness of the limitations that formalists have imposed upon language and life. If there is a universal grammar it is contextual, like meaning, whose structure inevitably conduces to symbolic expression. To reduce grammar to 'formal' parts is to reduce biology to the study of viruses.
The final stanza of a poem by Robert Frost, 'Stopping by the Woods on a
Snowy Evening,' reads as follows:
The woods are lovely, dark, and deep,
But I have promises to keep,
And miles to go before I sleep,
And miles to go before I sleep.
In a formal interpretation, the syntax and grammar of the last two lines are identical. But they are not. The context of repetition transforms these lines; their meaning is quite distinct because of that transposition. The first line in the sequence may refer to the literal distance the traveler in the woods has to go that evening, the second to a more symbolic expression of life in general: the miles of experience, and the final sleep. Like protein chains in DNA transcription, identical sequences mean very different things because of the context. The tension between the similarity and the differences is the essence of symbolism.
The organizational structure of these lines is both identical and very different, hence the earlier reference to fractals: the same formula, as it were, conduces to a very different picture in subsequent iterations that are themselves lovely, dark, and deep.
Alberts, Bruce, Dennis Bray, Julian Lewis, Martin Raff, Keith Roberts, and James D. Watson (1989) Molecular Biology of the Cell. 2d edition. New York: Garland Publishing.
The American Heritage Dictionary of Indo-European Roots, (1986) New York: Houghton Mifflin.
Arnold, Doug, Martin Atkinson,Jacques Durand, Calire Glover, and Louisa Sadler, (1989) Essays on Grammatical Theory and Universal Grammar. Oxford, Clarendon.
Bohr, Niels.(1928) The 1927 Como Lecture, in Nature 121: 13-22.
Bohr, N. in Atomic Physics and Human Knowledge. New York: J. Wiley and Sons
Chomsky, N. (1965) Aspects of the The Theory of Syntax. Cambridge, MA. MIT Press.
Chomsky, N.(1988) Language and the Problems of Knowledge. Cambridge, Mass.
Claiborne, R.(1989) The Roots of English. New York: Times Books.
Cherniak, C.(1986) Minimal Rationality. Cambridge, MA: MIT Press.
Churchland, Patricia S. Language, thought, and information processing. Nous 14: 147-170.
Churchland, Particia. (1986) Neurophilosophy. Toward a Unified Science of the Mind-Brain. Cambridge, MA: MIT Press.
Cooper, Jack. Floyd Bloom, and Robert Roth (1991). The Biochemical Basis of
Neuropharmacology. 6th ed. New York: Oxford Univ. Press.
Cytowic, R.E.(1989) Synethesia: A Union of the Senses. New York: Springer Verlag.
Damasio, A.R. and H. Damasio (1989). Lesion Analysis in Neuropsychology. Oxford: Oxford Univ Press.
Dennet, D.C. , Consciousness Explained. Little, Brown and Company. Canada.
Edelman, G. M. The Remembered present.: A Biological Theory of Consciousness. New York: Basic Books, 1989.
Edelman, G.M. Bright Air, Brilliant Fire. On the Matter of the Mind. New York: Basic Books, 1993.
Givon, T. (1979) On Understanding Grammar. New York: Academic Press.
Hawkins, J.A. (1983) Word Order Universals. New York: Academic Press.
Kandel, E., James Schwartz, and Thomas Jessell. (1991) Principles of Neural Science. 3rd ed. New York: Elsevier Science Publishing.
Lakoff, George , Women, Fire, and Dangerous Things - What Categories Reveal about the Mind, The University of Chicago Press, Chicago.
Lamham, Richard A. (1969) A Handlist of Rhetorical Terms. Berkeley: Univ. of
Langton, Christopher G. (1986). Studying artifical life with cellular automata.
Physica 22D: 120-149.
MacCormac, E. (1985), A Cognitive Theory of Metaphor. MIT Press, Cambridge, Massachusetts.
McDonald, J.F. (1992). Macroevolution and retroviral elements. Bioscience 40
no. 3 183-191.
Miller, J.(1985) Sematics and Syntax. Cambridge: Cambridge Univ. Press.
Ong, Walter. (1977) Interfaces of the word: studies in the evolution of consciousness and culture. Ithaca, N.Y.: Cornell Univ. Press.
Pask, Gordon (1975). Conversation, Cognition, and Learning. Elsevier.
Pattee, H.H. (1979) The complimentarity Principle and the origin of & nbsp; macromolecular information Biosystems 11: 217-226.
Pattee, H.H. (1982) Cell Psychology: An Evolutionary Approach to the Symbol Matter Problem. Cognition and Brain Theory 5: 325-341.
Pellionisz, A. and R. Llinas. (1979) Brain modeling by tensor network theory and computer simulation. The cerebellum: Distributed processor for predictive coordination. Neuroscience 4: 323-348.
Pellionisz, A. and R. Llinas (1985). Tensor network theory of the metaorganization of functional geometries in the central nervous system. Neuroscience 16: 245-273.
Pinker, S. and J. Mehler, eds. (1988) Connections and Symbols. Cambridge: MIT Press.
Pollack, R. (1994) Signs of Life. The Language and Meanings of DNA. Boston: Houghton- Mifflin.
Puttenham, G. The Arte of English Poesie (1589). ed. Gl.Willcock and A.
Walker. Cambridge: Cambridge Univ. Press, 1936.
Quintilian (1977). Institutiones oratoriae. Oxford: Clarendon Press.Rosch, Eleanor (1975). Cognitive Representations of Semantic Categories." Journal of Experimental Psychology, 104, 192-233.
Rosenfield, Isreal. (1992) The Strange, Familiar, and Forgotten. New York:
Sacks, Sheldon, ed. (1978) On Metaphor. Chicago: Univ. of Chicago Press.
Stein, Barry E. and M. Alex Meredith (1993). The Merging of the Senses.
Cambridge, MA: MIT Press.
Transposable Elements and Evolution. special edition of Genetica 86 nos. 1-3,
Van Fraasen, Bas C. (1980). The Scientific Image. Oxford: Clarendon Press.
Van Fraasen, Bas and C.A. Hooker (1976). in Foundations of probablility
theory, statistical Inference and Statistical Theories of Science. vol. 3. ed.
W.L. Harper and C.A. Hooker. 221-241. Dordrecht, Holland: Reidel.