Linguistics is arguably the most hotly contested property in the academic realm. It is soaked with the blood of poets, theologians, philosophers, philologists, psychologists, biologists, and neurologists, along with whatever blood can be got out of grammarians.
Russ Rymer in The New Yorker
Light, heat, gravity, chemical affinity, human will, have this in common, that they redistribute force. Their unit of process can be represented as: term from which, transference of force, term to which. If we regard this transference as the conscious or unconscious act of an agent we can translate the diagram into: agent, act, object.
Ernest Fenollosa, The Chinese Written Character as a Medium for Poetry
THE DOMINANT CONTEMPORARY THEORY of the origin of language proposes that genetic change produced genetic instructions for building a special module for grammar in the human brain. Before genetic specialization for grammar, people had no grammar at all: no grammatical speech, no parsing of grammar, no concept of grammar. To be sure, they communicated (birds and bees communicate), but their communication was totally ungrammatical. It was not language. This grammar module was autonomous: it borrowed no structure or processes from any other capacities like vision, spatial navigation, understanding of force dynamics, parable, and so on. Adherents of this theory—who form a large group of distinguished scholars that includes Noam Chomsky, Steven Pinker, and Paul Bloom—disagree only about which evolutionary mechanisms were responsible for the genetic specialization for grammar.
Naturally, it is a corollary of this theory that the development of language in any modern human child comes entirely from the autonomous grammar module in the child’s brain, which is built entirely from the special instructions in its genes. The language the child hears prompts it to shut down the parts of the language module it does not need.
I think this theory of the historical origin of language is wrong. A carefully adjusted version of it might not in principle be absolutely impossible, but at best it offers a hypothesis of desperate last resort: Since we cannot discover a straightforward way in which language might have arisen, let us postulate the mysterious origin of a special, autonomous black box that mysteriously does everything we need to explain language, including everything we don’t yet know we need.
If we reject the hypothesis that genetic specialization for grammar was the origin of language, what can we propose instead? Let us consider the possibility that parable was the origin of language, that parable preceded grammar.
Occam’s razor is a basic principle of theory building, named after the man who expressed it: Make no unnecessary hypotheses. We have seen that, independently of questions of grammar, we must concede that human beings have the mental capacities I call parable. Is it necessary to add to parable something new? Is it necessary to make the additional hypothesis that special autonomous instructions arose in human genetic material for building an autonomous black box in the brain that does the entire job? Not if we consider that parable already gives us what we need. Cognitive mechanisms whose existence we must grant independently of any analysis of grammar can account for the origin of grammar. The linguistic mind is a consequence and subcategory of the literary mind.
Stories have structure that human vocal sound—as sound, not language—does not have. Stories have objects and events, actors and movements, viewpoint and focus, image schemas and force dynamics, and so on. Roughly, parable takes structure from story and gives it to voice (or bodily signs in the case of sign language). Parable creates structure for voice by projecting structure from story. The structure it creates is grammar. Grammar results from the projection of story structure. Sentences come from stories by way of parable.
Parable draws on the full range of cognitive processes involved in story. Story involves spatiality, motor capacities, the sensory modalities (sight, hearing, touch, smell, taste) and submodalities, patterns that run across sensory modalities and submodalities, perceptual and conceptual categorization, image schemas, and our other basic cognitive instruments. Parable draws on all of this structure to create grammatical structure for vocal sound. Grammar, built from such structure, coheres with it.
Grammar arose in a community that already had parable. The members of that community used parable to project structure from story to create rudimentary grammatical structure for vocal sound.
Consider the following analogy. Imagine a community of people who have trained themselves in rudimentary martial arts. All members of the community have it. No genetic instruction for specialization in martial arts exists; the competence is assembled by directing to its use preexisting capacities of muscle control, balance, walking, vision, and so on, combined with arduous work to acquire it. But once it is acquired, it seems entirely natural, and, with a little practice, inevitable.
Now suppose that into this community a special infant is born with just a little genetic structure that helps it direct these preexisting capacities to this community’s martial artistry. The members of the community are better martial artists than the child and devote time to training and coaching the child, but the child has a secret edge. If the community is structured so that better martial artistry confers reproductive advantage, then the community provides an environment of evolutionary adaptiveness for the genetic change: the “martial artistry” trait is adaptive. This situation could plausibly give rise to a kind of genetic arms race in which each increment of further genetic specialization brings an increment of relative reproductive advantage. But martial artistry itself arose without genetic specialization for martial artistry.
Now imagine a community of people who use parable to create rudimentary grammatical structure for vocal sound. Everyone in this community develops story and projection, has voice, receives training from his parents, and is assimilated into the work of creating grammar through parable. Suppose that into this community a special infant is born with just a little genetic structure that helps it project story onto voice. The members of the community are better at rudimentary language than the child and devote time to training and coaching the child, but the child has a secret edge. If the community is structured so that greater facility with grammar confers reproductive advantage, then the community provides an environment of evolutionary adaptiveness for the genetic change: the “grammar” trait is adaptive. This situation could plausibly give rise to a kind of genetic arms race in which each increment of further genetic specialization brings an increment of relative reproductive advantage. But grammar itself arose without genetic instruction for grammar. It arose by parable.
There are basic abstract stories. A basic abstract story is projected to create a basic kind of grammatical structure. For example, the basic abstract story in which an animate agent performs a physical action that causes a physical object to move in a spatial direction is projected to create the grammatical structure we see in “John pushes the ball onto the court,” “David tosses the can into the yard,” and “Mary throws the stone over the fence.” The abstract narrative structure is projected to create the abstract grammatical structure. The abstract narrative structure includes an agent, an action, an object, and a direction. The abstract grammatical structure includes a noun phrase followed by a verb phrase followed by a noun phrase followed by a propositional phrase, with the first noun phrase as Subject and Agent and the second noun phrase as Direct Object and Patient.
The first abstract structure is conceptual and narrative. The second abstract structure is grammatical. If we think of these two structures as residing in two spaces, then there is a generic space that contains just the structure they share. This generic space is more abstract than either of them; its structure is not specifically conceptual or grammatical; it includes only elements, distinction of elements, some relations between elements, and so on.
The story in which Mary throws a stone and the story in which Bill flips a coin are different in nearly every specific detail, but we take them as sharing an abstract story structure. “Mary throws a stone” and “Bill flips a coin” are, as vocal sound, different in nearly every detail, but we take them as sharing an abstract grammatical structure. The abstract story structure and the abstract grammatical structure share generic structure.
Let us consider an example of the way in which story is projected to create grammar. Consider the small spatial stories in which Mary throws a stone, John pushes a ball, and David tosses a can. These small spatial stories are all instances of the same basic abstract story.
This basic abstract story has certain kinds of structure. One kind of structure it possesses is distinction of certain elements—Mary, the act of throwing, the stone. Notice that if we actually see (or imagine) Mary throwing a stone, we cannot distinguish the perception of throwing from the perception of Mary and the perception of the stone. Nonetheless, conceptually, in narrative imagining, we distinguish these three elements.
These three elements have category structure. Mary, for example, is placed into the category of animate agents; throwing is placed into the category of events; the stone is placed into the category of objects.
The story also has combinatorial structure. The distinguished elements of the story include Mary, the stone, the causal relationship between Mary and the throwing, the causal relationship between the throwing and the movement of the stone, the causal relationship between Mary and the movement of the stone, the event shape of the throwing, and our temporal viewpoint with respect to the throwing, all of which are combined as simultaneous: the act of throwing involves all of them at once. This combination has hierarchical structure—having a viewpoint on a story depends upon the existence of the story, for example.
In sum, the abstract story has certain kinds of structure: reliable distinction of elements, distribution of elements into categories, simultaneous combination, hierarchy, and so on. Other basic stories show recursive structure: if Paul catches the stone Mary threw, then one story (Mary throws a stone) feeds into a second story (Paul catches the stone).
Vocal sound itself—as sound, not language—does not have this structure. The elements of the story have a reliable structure of distinction but the sound “Mary throws a stone” is more or less a continuous stream that, if divided up at all, could be divided up any number of ways. The elements of the story have a reliable hierarchical structure of joined but conceptually distinguished elements (e.g., the event and the temporal viewpoint on the event) that the sound—again as sound, not language—does not mirror. The elements of the abstract story have category structure that the sound does not mirror: If Mary throws, John pushes, and David tosses, then Mary, John, and David belong to a category, but the sounds “Mary,” “John,” and “David” belong to no such reliable category. The causal structure in the abstract story has nothing to do with the causal structure of vocal sound. The temporal structure of vocal sound is always linear sequence but the temporal structure of this story involves highly complex simultaneity; other stories involve even more complicated temporal structure. In sum, story and vocal sound are two very different sorts of things. Story structure is projected to create structure for vocal sound that vocal sound does not intrinsically have.
The distinction of elements in the abstract story is projected to make “Mary,” “throws,” and “a stone” precisely distinct not as sound but as grammatical elements. The category structure in the abstract story is projected to vocal sound to put “Mary” and “a stone” into the same grammatical category—noun phrase. As sound, “Mary” and “a stone” share no reliable category, but as grammar, they do. The different roles of Mary and the stone in the story are projected to give “Mary” and “stone” different categories of grammatical relation (Subject versus Direct Object) and semantic role (Agent versus Patient). The structure of temporal foci and viewpoint in the abstract story is projected to give the sentence grammatical tense.
Abstract stories are projected to create abstract grammatical structures. The story in which Mary throws a stone is an instance of an abstract story, that abstract story is projected to create an abstract grammatical structure. The abstract story has indefinitely many instances; the corresponding abstract grammatical construction has indefinitely many instances. Mary throws a stone, John pushes a ball, and David tosses a can are all instances of the abstract story; “Mary throws a stone,” “John pushes a ball,” and “David tosses a can” are all instances of the corresponding abstract grammatical structure.
The abstract story, in one mental space, has conceptual structure. The abstract grammatical structure, in a second mental space, has grammatical structure. The very abstract structure they share resides in a generic space. It may sound odd to say that we blend the abstract story with the abstract grammar, but nothing is more common: In grade school, we are taught that a Noun is a person, place, or thing, a Verb is an action or event, and so on. Of course, a Noun is certainly not a person, place, or thing, and a Verb is certainly not an action or an event. Nouns and Verbs are grammatical; people, places, things, actions, and events are not grammatical. But in the blend, we join them. “Mary throws a stone” seems to be both a grammatical sentence and a story. Sentences are stories. Drawing inspiration from work done by Charles Fillmore and Paul Kay and later work done by Adele Goldberg, I call a blend of story structure and grammatical structure a grammatical construction. In a construction, certain story structures go with certain grammatical structures. When we want to tell that story, we use that grammar. When someone uses that grammar, it prompts us to think of that story.
Let us take a broad look at some of the principles of this “parabolic” view of language. The projection of story structure to create grammatical structure for vocal sound is not one projection but indefinitely many. There are indefinitely many specific projections of story to voice; the vasdy complex network of all these specific projections is “the projection of story onto voice to create grammar.” As narrative structure is not one thing but rather a mental activity that coordinates very many stories at many levels, so grammatical structure is not one thing but rather a mental activity that coordinates very many grammatical structures at many levels. The origin of grammar is the establishment of a dynamic coordinated complex of different but related grammatical structures that arise from projections of story structure. Grammar is not the beginning point of language: parable is the beginning point of language. Grammar arises originally from conceptual operations. Rudimentary grammar is a repertoire of related grammatical constructions established through parable. The backbone of any language consists of grammatical constructions that arise by projection from basic abstract stories.
Story and grammar have similar structure because grammar comes from story through parable. Let us consider some of the similarities between them.
Story depends upon constructing something rather than nothing. A reportable story is distinguished from its assumed and unreportable background. It is impossible for us to look at the world and not to see reportable stories distinguished from background, even though distinguishing in this fashion is hard to justify from the point of view of physics and biology. If we look out of the window and someone asks us what is out there, we can reply “Nothing” and mean it, so long as what we are looking at seems like background: a tree and a lawn on a quiet day. But if a lightning bolt strikes the tree, it looks very different, and becomes reportable. We believe that the same laws of physics and biology hold for both scenes and that in a scientific sense a great deal is going on in both scenes. Yet the lightning strike looks like a little story and the other scene looks like background, nothing remarkable.
This distinction of story as opposed to no story projects to the distinction of speech as opposed to no speech. Silence is part of grammar and is the counterpart of no story. We grant to certain elements of our perceptual experience special distinguished status. We think of things as happening against a background. Analogously, we think of speech as happening against a background of silence. The silence reports the background; the speech reports whatever is distinguished against the background. This is not always the case, of course: modern languages have developed many highly intricate instruments that go far beyond the basic grammatical constructions for basic stories. But radically, we report that the lightning struck the tree or that one person hugged another or that the rains flooded the tunnel or that the branch waved back and forth in the heavy wind. Reporting a nonevent is less common, unless the event was expected, in which case the nonevent is itself conceived of as an event. The form of reporting nonevents—The lightning did not strike the tree”—is also much less prototypical. It would be odd to make this report if the event had not been expected; the negation pays homage to the story that did not happen rather than the nonevent that did.
Narrative imagining partitions and categorizes wholes into related elements, as when the small spatial story of Mary throwing a stone is perceived and categorized as involving different elements in different categories that stand in narrative relationships. Narrative partitioning, categorizing, and relating project to create grammatical partitioning, categorizing, and relating. The result is grammatical categories and grammatical relations.
Story groups elements. For example, the image schema path-to-goal groups the elements in the motion of the stone toward the window into a single unit. Story grouping projects to create grammatical grouping: “Toward the window” is a grammatical group. In English, this grammatical grouping is not simply a matter of sound adjacency. “Mary threw the rock toward the—if I’m not mistaken—window” still has “toward the window” as a grammatical group. Nonadjacency in grammatical grouping is even more evident in a language like Latin.
Narrative imagining combines finite elements into infinitely many products—Mary, John, throws, catches, pushes, flips, stone, ball, and coin, as conceptual elements of narrative imagining, combine into a great number of particular stories: Mary throws the stone, John throws the stone, Mary throws the ball, John catches the ball, John flips the coin, Mary pushes the coin, and so on. This structural property of story is projected to create a structural property of grammar: Finite grammatical elements combine into infinitely many products. “Mary,” “John,” “stone,” and “ball,” as grammatical elements, combine into a great many grammatical products: “Mary throws the stone,” “John throws the stone,” “Mary throws the ball,” and so on.
Narrative imagining is exceptionally adept at nesting—putting one story inside another. Story nesting is projected to create grammatical nesting: “The house that Jack built fell down.” Literary stories like the Odyssey or The Thousand and One Nights have stories nested to several layers, but it is hard to track all these layers at once, a fact Shahrazad exploits repeatedly to great rhetorical effect. As has often been observed, the same is true of grammatical nesting: “The house that Jack whom my aunt who came from Germany named built fell down” is hard to track
Different basic stories are organized in a network; they are not independent of each other. They share structure. For example, the story categories Agent and Action show up in very many different basic stories. Just so, the different grammatical constructions that come from different basic abstract stories are organized in a network. They are not independent of each other. For example, the basic grammatical categories Noun and Verb show up in very many different basic grammatical constructions. Basic kinds of elements in stories have status of their own: Agent is a unit in the conceptual network of story. Just so, basic kinds of elements in grammatical structures have status of their own: Noun and Verb are themselves constructions in the network of grammar. Narrative imagining works with a network of related story structures; grammar works with a network of related grammatical constructions. Although modern language has developed many tricks, the original principle of networking persists: modern grammar is a dynamic and adaptive network of constructions, new ones evolving over time as basic ones continue to do their job.
The projection of basic stories accounts for the existence of corresponding basic grammatical constructions. That is a small beginning. Now we add a crucial development. We saw repeatedly in the early chapters of this book that the structure of a basic story can be projected onto other conceptual targets. For example, the structure of an agent beating an object can be projected onto the story of what the wind does to the sailor; the structure of taking an object away can be projected onto the story of Death. The target of such a projection has, by parable, structure that matches the source. Suppose a basic story projects to both a target story and a grammatical construction. Then the structure of the grammatical construction matches the structure of the target story. In that case, the grammatical construction can be used to express not only the basic story but also the target story. In sum, grammatical constructions that represent a basic story represent any conceptual target structured by projection from that story. We say both “Mary threw the stone out the window” and “Mary threw the job out the window.” The first is a small spatial story. The second is not at all spatial. The nonspatial story has structure projected from the spatial story. The grammatical construction has structure projected from the spatial story. Therefore, the grammatical construction represents both the spatial story and the nonspatial story. This phenomenon—the same grammatical construction for expressing stories of profoundly different features—is such an indispensable part of our thought and language as to seem unproblematic. But something this profound requires an explanation, and the explanation is parable.
A great range of things that are not prototypical objects, events, agents, or actions in a story can be conceived by projection as if they were. For example, times are not moving objects, but by parable we can think of them so, and this target can then be expressed by the grammatical construction that corresponds to the source: We say “The deadline is approaching” in the same way we say “The car is approaching.”
The event of thinking is a process, not an object or an agent, but it can be conceived as an object-agent through parable. Conceived in this way, the target can be expressed by the grammatical constructions that correspond to the source story: We say “His thinking is moving in the wrong direction” just as we say “His truck is moving in the wrong direction.” We say “I cannot grasp the idea” just as we say “I cannot grasp the handle.” We say “I am turning it over in my mind” just as we say “I am turning it over in the pot.” We say “I grasped the idea” just as we say “I grasped the object.” We say “He accepted the explanation” just as we say “He accepted the gift.”
There are other ways in which rudimentary language can be extended. A grammatical construction like Subject Noun Phrase that arose by projection from one part of story structure like Agent can be exploited to express related parts of story structure like Instrument: “John broke the stick” has a grammatical construction that can be exploited to say, “The stone broke the stick.”
Where the patient in a small spatial story seems itself active in an activity induced by an agent, we may exploit the Subject position to express not the principal agent but rather the “active” patient: “The stick broke.”
And so on. For parable to be the root of language means that it supplied the beginning of language. That beginning was developed and was exploited for the great additional range of communicative purposes.
Let us take, as an example of how grammar comes from parable, some features of tense in English. I pick tense as an illustration because although it has been studied explicitly, intensively, and intelligently for at least two and a half millennia, and although the technical literature on tense as a facet of many different languages is voluminous, scholars of tense agree that tense is not well understood as a grammatical phenomenon. Robert Binnick begins his encyclopedic 1991 work, Time and the Verb, “It is no contradiction to say that we know a very great deal about tense, but understand it little It has been difficult even to know how much we do understand it, for confusing as discussions of the tenses of various languages may be, the scholarly literature concerning tense in general is, if anything, even more confusing.” It seems reasonably clear that any general explanation of tense finally accepted will look substantially different from what we have now.
We explored in chapter 7 the basic human story of a person recognizing a story. In this general story, there is a recognizing agent who has a single spatial focus and a single spatial viewpoint. We saw that projecting the story of perception in space with spatial focus and spatial viewpoint onto the story of perception in time gives us an agent with a temporal focus and temporal viewpoint. Someone recognizing a story in time has, by means of this projection, a particular temporal focus and a particular temporal viewpoint.
In spatial perception, there are alternative foci and viewpoints. In temporal perception, there are alternative temporal foci and temporal viewpoints. This structure of temporal foci and viewpoints is projected to create a rudimentary grammatical system of tense. The projection of certain relations between temporal focus and temporal viewpoint creates certain constructions of tense. In English, basically, the grammatical construction of present tense corresponds to the narrative category in which temporal focus and viewpoint are the same; the grammatical construction of past tense corresponds to the narrative category in which the focus precedes the viewpoint; the grammatical construction of future tense (or technically, of verb phrase constructions that we commonly say signify “future tense”) corresponds to the narrative category in which the viewpoint precedes the focus. A relative tense such as we find in “I will have run” corresponds to a complex narrative structure in which viewpoint precedes focus, but the focus itself contains a viewpoint that is preceded by its own different focus. Technically, spaces A and C precede space B; C has the event of running; A has a speaker considering B; B has an agent identical to the speaker and the runner and who is considering space C. So A is a viewpoint on B as a focus and B is a viewpoint on C as a focus. Tense lets us express such a structure of temporal viewpoints and foci as “I will have run.”
Often we are told—or have reason to suspect, or assume from background knowledge—that the temporal viewpoint is identified with the actual moment of speaking, but this is an independent circumstance. Although it is the default circumstance, it is not fixed and does not have to be used. In story, the identification is optional; it is therefore optional in tense. Linguists often regard an utterance in English that makes no use of this identification as a problem, an exception that requires special explanation. I offer a different view. In the parabolic view of tense, present tense arises as the counterpart of the narrative structure in which temporal focus and temporal viewpoint belong to the same space. Only in the default case does viewpoint also correspond to the moment of speaking. Cases that do not fit the default scenario are not problems or exceptions. Let us consider some examples.
I will begin with examples that make the nonidentification of viewpoint and moment of speaking very obvious—they sound odd out of context, but people said them.
In the first example, the setting is a veranda; an old man is sitting; his old wife comes out, and, after a pause, sits down beside him; the silence grows more nervous; at last, the old man says, “It was a warm night on the banks of the intercoastal waterway. He sat alone, until she finally appeared and sat beside him. She seemed distracted and unhappy. At last, she announced—” Here, the man telling the story gestures to his wife to announce whatever it is she has to announce; the man and woman in the story are the old man and old woman on the veranda. In this example, the past tense tells us that the focus lies behind the viewpoint. From context, we know that the focus is the moment of speaking. The viewpoint is not the moment of speaking. Instead, the moment of speaking precedes the viewpoint.
Consider another lively real example with similar structure. A wife and husband are traveling in Greece; at night, they defend against mosquitos by covering the hotel window with netting; kounoupi is Greek for “mosquito.”
I am going to clean up, take a shower, lay down my naked body.
What about the kounoupi netting?
My husband took care of that while I was in the shower.
The past tense tells us the focus precedes the viewpoint: The focus is the space in which the husband puts up the kounoupi netting; the viewpoint is sometime after the focus; but we know from context that the moment of speaking precedes the viewpoint (as well as the focus).
Now consider idiomatic examples that have similar structure. Two travelers ask the train attendant to tell them the destinations of each of five trains that are sitting at platforms; the conductor tells them; the travelers think it over, and one of them asks, “That train, where did it go?” In this case, the focus precedes the viewpoint but we know from context that the moment of speaking precedes both of them. Later, one of the travelers can start to hop on the wrong train and the other can say, “Wait! That train went to Nice!” Consider also:
Which of these things do you want to give to charity?
That one goes, that one goes, that one goes, that one stays. When?
The moment the charity truck shows up, they go.
Here, viewpoint and focus are the same and we know from context that the moment of speaking precedes both. Consider the setting in which Jane comes home and gripes at Mary that the dishes are not done. Mary says, “Today, I will get up early to run a million stupid errands, clean the bathroom, shop for food, entertain Jane’s charming mother for lunch, and prepare dinner. Jane will come home and then Jane willgripe about the dishes. I will shoot her between the eyes, and the grand jury will decline to indict.” Here, viewpoint precedes focus but we know from context that the moment of speaking is bracketed by the focus.
There are many other recognized cases in which viewpoint does not correspond to moment of speaking. In “historic present,” focus is viewpoint but we know that both precede moment of speaking: “It’s the late fifteenth century: Columbus has just anchored and is being rowed to the shore. He sees palm trees and a sandy beach.” “It’s the dawn of humankind: a smaller than average primate picks up a leg bone and smashes the head of his antagonist.”
In “futuric present,” focus is viewpoint but we know that moment of speaking precedes both:
So, what do you say tomorrow when you ask for a raise?
I go in, she asks me about the financing, I tell her it’s finished, she beams and compliments me, and I say …
There are also cases in which viewpoint precedes focus and focus precedes (often by centuries) moment of speaking: “The Theran explosion will wipe out the Minoans, the Achaeans will burn Troy, Alexander will make the city-state obsolete, and Rome will erase almost everybody. But the archaeological record of everyday implements will survive for us to unearth and explain.”
There are many such examples. They are not exceptions, but instances of the general case in which narrative imagining includes relations of temporal focus and viewpoint that are projected to create rudimentary grammatical constructions of tense. In the default instance, viewpoint may correspond additionally to moment of speaking. The existence of the default is understandable: Although we may shift temporal viewpoint in imagination, viewpoint and moment of speaking occur simultaneously in our perceptual present. But this additional default association is not the general case, and other associations are not exceptions.
This is not the standard way of analyzing absolute tenses. Typically, explanations of tense begin from the view that tense expresses the temporal relationship between the event reported and the moment of speech: past tense for events before the moment of speech, present tense for events overlapping the moment of speech, future tense for events after the moment of speech. This is the beginning frame used by Bernard Comrie in Tense: “past time reference is the basic meaning of the past tense”; “As far as the present tense is concerned, in its basic meaning it invariably locates a situation at the present moment.” It is equivalently the standard beginning frame for analyses in the tradition of Hans Reichenbach’s influential Elements of Symbolic Logic.
This beginning frame is often stated rather technically, as in John Dinsmore’s formulation in “The Logic and Functions of the English Past and Perfect,” in which he proposes that the following rule expresses the meaning of English Past:
(Past) For any time T and sentence S, at(T, pa(S)) iff at(T, S) and T < now.
Buried in this formalism is the claim that the past of a sentence is true of a certain time if and only if the sentence is true of that time and that time precedes the moment of speaking. This does not fit “My husband took care of that while I was in the shower”: yes, the present-tense version of the sentence is true of the narrative mental space of taking the shower; yes, the past-tense version is true of that mental space as viewed from the viewpoint; yes, the time of that mental space precedes the time of the viewpoint; but no, the the time of that mental space does not precede the moment of speaking. On the contrary, moment of speaking precedes focus and focus precedes viewpoint.
Everyone recognizes that there are such exceptions. The typical way to handle these exceptions is ad hoc, one at a time, through local contrivance. In As Time Goes By, Norbert Hornstein observes that the principal exception occurs when a text “establishes a date that then acts as the anchor for the interpretation of the tenses used,” as in “It was 1812, just before the Battle of Borodino. The anticipation of the coming struggle is palpable.”
But “My husband took care of that while I was in the shower” does not explicitly establish a time that acts as the viewpoint. Rather, the tense itself indicates that the narrative mental space containing the focal event precedes the viewpoint. The relationship of this viewpoint to the moment of speaking is not part of the grammar but is rather something that we establish pragmatically. Our default pragmatic assumption may be that the viewpoint is the moment of speaking, but, importantly, that default assumption does not have to be used, as all such examples show.
The view of tense as arising from the projection of temporal viewpoint and focus requires no extra machinery beyond parable. It also saves us from ad hoc explanations. For example, how are we to explain a phrase like “John goes to work at eight o’clock every day”? Bernard Comrie, in an effort to accommodate such examples while still requiring the present tense to situate the event in the present moment, writes, “A certain property (namely, going to work at eight o’clock every day) is assigned to John, and this property is of course true of John even if at the moment he happens not to be on his way to work. In other words, the habit does hold at the present moment, and that is why the present tense is in principle an appropriate tense to use in describing this habitual situation.” Aside from the fact that it is hard to see how going to work in the morning is a habit that holds at other times than morning, or that holds when one is in fact not going to work at all—as during vacations, hospital stays, and so on—this explanation will not handle a case such as the train attendant’s saying “That train goes to Paris” of a train that will leave for Paris tomorrow morning from a station from which no train has previously departed for Paris and no train ever will again; nor will this explanation work for “Curtis rents the boat” said as part of the presentation of a plan for a nonhabitual future.
Comrie has the similar difficulty for the past tense—as when the waiter asks the Russian equivalent of “Who received the goulash soup?” or says in English, “Who had the roast beef sandwich?” In these cases, the tense is past but the focal event of receiving or having lies in the future of the moment of speaking. There are harder cases still. In several languages, like Russian, Comrie notes, the past can be used for imminent future events, as in “I left” when one is about to leave. English and French have a similar use for the present, as in “I am coming” or “Jy suis.” Comrie’s expedient, a common one, is to regard such “nonliteral” uses of tense as constituting exceptions that lie outside the scope of his analysis: “Rather, it seems that such uses of the past should simply be treated as exceptions.”
Another “exception” is an extremely common use of the present perfect. Theories of tense often regard the present perfect as marking that the time of the event precedes the moment of speaking. But it can be used for events that one expects to happen, as in “John has won the race,” said when John has not yet crossed the finish line, or “Clinton’s won,” said of the candidate before the election. A common explanation for this “exception” claims that when the preconditions of an event’s taking place have been fulfilled, then it can be marked as preceding the moment of speaking even though the event is in the future of the moment of speaking.
On the theory that tense arises as a grammatical structure by projection of viewpoint and focus, none of these cases has an exceptional mechanism, they simply do not choose to connect viewpoint with the moment of speaking. “John goes to work every morning” cues us to take a viewpoint on a story of John going to work. The time of the narrative mental space is morning, and the viewpoint taken is also morning, and we know pragmatically the relation between the moment of speaking and “morning.” The train attendant takes a temporal viewpoint on a story of the train’s going to Paris that is the same as his temporal focus, and we know pragmatically the relation of that viewpoint to the moment of speaking. The viewpoint on the story of the customers’ receiving their food lies ahead of that space; the viewpoint on the story in which the person who says “I am coming” is coming is the same as the focus; the viewpoint on the story in which John wins the race lies ahead of that narrative mental space. In all these cases, we know the relation between the temporal viewpoint and the moment of speaking pragmatically; the default pragmatic identification of temporal viewpoint and moment of speaking is simply not used.
The reasons for taking such viewpoints are not always hard to compute. For example, in the case of “John has won the race,” said when John has not yet crossed the finish line, I would say that putting the event in the past of the viewpoint may serve to indicate a confidence in the event because presumably the past is fixed. Alternative analyses according to which such cases are special exceptions requiring special rules often adduce exactly these reasons. An analysis of “John has won the race” as an exception might say that the special rule is this: The “perfect” marking of the event as anterior to the moment of speaking but relevant to the moment of speaking can be used when preconditions of the event are fulfilled at the moment of speaking. These two different explanations share most of their substance. But there is a large difference in the way they frame the analysis of tense. In a view of tense as a grammatical system that arose by projection from narrative structure, the “nonliteral” cases are not essentially exceptions to the system of grammatical constructions for tense, and, more importantly, tense can have arisen through machinery whose existence we must already concede: parable.
Even a tough case like “Cows eat grass” does not look unusual on the parabolic explanation: There is a story of typicality in which everything is doing or being what it typically does or is; the time of that story is eternal; the viewpoint on that space is neither before nor after it; and we know pragmatically the relation of that story to our mental space of our own present reality or to the speaker’s mental space of the speaker’s reality.
The world does not occur automatically as narrative; even less does it occur automatically as a unique narrative. But it often seems to us as if it does. It may seem as if seeing reality as having a time line with a fixed viewpoint called “the present” requires no narrative imagining. But we are free to choose any other temporal viewpoint. Tense, which arises as a grammatical system by projection from the system of temporal viewpoint and focus, is an instrument for indicating our choice.
The projection of narrative structure gives rise to rudimentary language, and once rudimentary language is set up, it is extensible beyond the projections that gave it rise. Narrative structure—and, more generally, conceptual structure—is incomparably more complicated than grammatical structure. Once grammatical structure is established by projection of narrative structure, it can be adapted to express vast ranges of conceptual structure beyond the structure that gave it rise. One way to adapt a grammatical construction is through metonymy: replacing an element with something conceptually associated with that element. For example, once “verb phrase” is set up as a grammatical construction by projection from the category of events and actions in stories, the verb phrase can be extended to accommodate not only events but also elements related to events. We say, “Mary hit the window,” but also “Mary broke the window,” because the window’s breaking is metonymically associated (as causal result) with the action Mary performed to break it, and “Mary stoned the window,” because the stone is metonymically associated (as instrument) with that action. Different constructions permit different metonymies.
A second way in which a grammatical construction is adapted is by pressuring it to accommodate other conceptual structures beyond those that gave it rise. For example, resembles is not a prototypical event and John resembles Joe is not a prototypical story, but the preexisting grammatical structure can be adapted to serve the purpose of expressing this situation in a nonprototypical sentence. Snow is white is not a narrative but rather a subset of information that might be relevant in a narrative, yet the grammatical structure created by parable can be adapted to serve such a propositional expression.
The view that language arises through parable in no way involves a claim that every construction or even most constructions in a speaker’s system of grammatical constructions can be explained as arising through projection of narrative structure, nor does it imply that we will not find an amazing range of uses for language once language is set up and begins to evolve, nor does it imply that all grammatical constructions are isomorphs of small stories, nor does it by any means entail the claim that the only thing language can express is narrative structure. Rather, we are seeking to explain how language could arise without additional machinery, that is, without postulating hypothetical agencies and mechanisms that, conveniently, have all the properties needed to solve the problem but for whose existence we have no independent evidence. A special mechanism to do just what we need can always be conjectured in science, but such a conjecture is always suspect, a deus ex machina, and is to be introduced only as a last resort. We are not driven to that last resort. Rudimentary language arose through parable; after arising through parable, it was extended.
The view of language as arising through parable would allow for diversity of languages and grammars. Details of narrative structure vary from culture to culture and even person to person: what is universal is not all the specific narrative structures but rather stability of basic abstract stories. All cultures have stable repertoires of basic abstract stories. Some of them vary culture to culture. An accusative language (which gives Subject and Agent one grammatical coding and Patient a distinguished grammatical coding) and an ergative language (which gives intransitive Subject and Patient one grammatical coding and transitive Subject and Agent a distinguished grammatical coding) may seem to be informed by different kinds of stories, but each will seem to have a stable repertoire of basic abstract stories.
Projection is widely variable in the actual structures it projects and the ways in which it projects them. Even when two different languages project the same basic abstract story, and thereby give rise to similar grammatical constructions, the details will often be strikingly different. In English, “John loves Mary” and “Mary loves John” have different meanings, but in Latin “Iohannes amat Mariam” and “Mariam amat Iohannes” mean the same thing. In both languages, narrative relations project to create grammatical structure, but in English the result is a grammatical structure of linear word order, while in Latin it is a grammatical structure of case endings on nouns. Similarly, English has one set of constructions for representing causal narrative chains (“I make Paul eat”) but French has a different and highly intricate set of double-verb causative constructions (“Je fait manger Paul”) that English lacks.
Grammar arises as a dynamic system of projections of story structures to create a dynamic system of grammatical constructions that then can be adapted dynamically in many ways. It is likely not only that different speakers in the same linguistic community will have somewhat different grammars but even that the same person will have somewhat different grammars at different times. What matters for language as a communicative device is not that the members of a linguistic community all have the identical dynamic system of grammar but rather that the external products of grammar are perceived as fulfilling the relevant communicative needs in local situations. When they are not so perceived, then objections, corrections, and negotiations can take place to tune the systems of the conversants. On the parabolic view, grammar is a kind of dynamic repertoire, in much the same way that perceptual and conceptual categories are repertoires. The repertoires need not be identical universally or even in a community, but they do need to be stable and effective.
The level at which the projection takes place is the level of constructions above individual lexical items. A particular word like “dog” is not the projection of the concept dog. Rather, grammatical constructions and systems of constructions like Noun Phrase, Verb Phrase, Subject, Object, Agent, Patient, Present Tense, Progressive Aspect, Caused-Motion, Ditransitive (for example, “He baked me a cake”), Resultative (for example, “She hammered it flat”) and so on arise from projection of systems of narrative structure. It is grammatical structure, not single lexical items, that arises from the projection of story.
There is a standard objection to any view that grammar arises from conceptual structure: we find in grammar various extremely odd formal quirks that seem to have no conceptual counterpart or functional explanation; therefore, so the logic goes, there must be an independent (innate) grammatical mechanism to give us 100 percent of grammar, including the quirks. But if grammar arises through parable, then grammar is based on everything involved in narrative imagining. It is based on all the sensory modalities and submodalities, on motor capacities, and on perceptual and conceptual categorization, which result in abstract structures like image schemas and dynamic integrative connections across different distributed activities in the brain. If it seems plausible that a genetically instructed grammar module would have quirks, we must a fortiori grant that the neurobiological and cognitive activity involved in all these other poorly understood systems would also have quirks—evolution is so notorious for producing quirks that the existence of quirks is a good test for separating evolved systems from rationally designed systems. Projection could project versions of those quirks when it gives rise to language. There is ample opportunity for the creation of quirks in grammar by projection from all of the systems underlying narrative imagining, without having to hypothesize an extra mechanism just to get the quirks.
There is also the possibility that some of the quirkiness may be arbitrary but necessary for consistency or efficiency in the system. Consider the analogy to copy text: We usually write to express meaning, and when we do so, it does not matterwhether we write dates in the form “March 24,1954” or “24 March 1954.” Such matters are largely arbitrary. Accordingly, we resort to manuals of style for consistent decisions on arbitrary matters. There might be a great deal in grammar that works one way as opposed to another for the sake of essentially formal consistency. There is room in linguistic analysis for both parable and arbitrary formal patterns, regardless of the question of the existence of a genetically instructed grammar module in the brain.
Vast and complicated as language is, it is small compared to conceptual structure, communicative purposes, and local construction of meaning. Conceptual activity and communicative situation place a great range of constraints on language to operate effectively and efficiently, and yet to stay relatively small and manageable. As grammar evolves, it must satisfy many and varied purposes and meet many and varied constraints. Some of the pressure on grammar comes from conceptual activity, some from communicative purpose, and some from problems of internal organization of the system of grammatical constructions. A consistent and workable solution to the problem posed by all these conflicting pressures is the grammar of a language, which will not look uniformly like a set of local and simple pairings between grammatical structure and story structure. Plenty of strange opportunistic tricks, many of them ad hoc, are to be expected, with quirks as one of their manifestations.
On the other hand, there is nothing in the view of grammar as arising from the projection of story that in principle rules out the evolution of specialization for the projection of narrative structure to the particular target of voice, and therefore nothing to say that quirks in grammar as we know it could not arise from genetic specialization. But the specialization is not in principle needed to get quirks.
Linguistics typically discusses grammar at the level of the sentence. The theory of grammar as arising from parable invites us to ask whether the structure of units of discourse higher than the sentence might share structure with sentential grammatical constructions. Grammatical constructions are units of discourse; higher units of discourse contain them. A handbook like Style—Joseph Williams’s summary of his collaboration with Gregory Colomb—which is dedicated to helping writers make their prose easier for readers to parse, tells writers that a major improvement can be achieved simply by making the subjects of clauses correspond to “the cast of characters” in the story and making the verbs that go with those subjects correspond to “the crucial actions those characters are part of.” Williams and Colomb observe that readers are more likely to find prose clear and direct when it seems to present characters and their actions, thereby telling stories. What Williams and Colomb do not observe is that such prose is clearer exactly because the grammar arises from those basic stories. Readers naturally expect the stories being told to line up with the structure of the grammatical constructions being used. They read prose more easily when it follows such an alignment. Williams and Colomb recognize that many things that are not actually characters or actions are understood as such by projection, and they frame their advice accordingly. They advise those who want to write clear prose to change passages like the first version below into passages like the second version:
Because the intellectual foundations of evolution are the same as so many other scientific theories, the falsification of their foundations would be necessary for the replacement of evolutionary theory with creationism.
In contrast to creationism, the theory of evolution shares its intellectual foundations with many other theories. As a result, creationism will displace evolutionary theory only when it can first prove that the foundations of all those other theories are false.
The cast of “characters” in the second version includes creationism, evolutionary theory, and other theories. The actions include sharing, displacing, and proving. The second version presents characters in the story and the actions in which they are involved. The first version does not. The second version is easier to read. In my words, the grammar carries expectations about narrative structure because it is created by projection from narrative structure; the second sentence meets the inevitable expectations of correspondence better than the first.
According to Williams and Colomb, readers will regard a paragraph as more cohesive if the topics of its sentences present consistent information—which usually means referring to the characters in the story. Such paragraphs will also be more coherent if their sentences’ topics present a consistent “point of view.” In my terms, the larger grammar of the paragraph derives from the conceptual structure of a story, in which characters are involved in actions. Readers therefore expect the string of topics from sentence to sentence to connect the main characters, and expect one of those characters to dominate this topic string so as to present the scenario from a consistent viewpoint.
If this is right, then the prototypical grammars of clauses, sentences, and texts all derive from the same source: parable.
Everyone agrees that acquiring language helps an infant develop mentally by making it possible for the infant to express thought and to understand the expression of other people’s thought. If grammar arises through parable, acquiring language could help the infant develop mentally in an additional, altogether different way. A grammatical structure corresponds to a story structure; the two structures are blended in a grammatical construction. Under parable, there can also be projection back from the blend to input spaces. The development of grammatical constructions could therefore reinforce the learning of story structure. The learning of a language may quite literally change the neurobiology of the infant in ways that are influential over cognition. This creates the intriguing possibility that speech and writing could be ways for the brain of one person to exert biological influence upon the brain of another person: thinking may be affected abidingly by experience with language.
If we use the old metaphoric conception of the brain as an agent who “deals with” language or as a container that for a moment “holds” language while examining it for storage or discard, then it is natural to think of the biology of the brain as unchanged by its dealings with language. But if we use instead the conception of the brain as an active and plastic biological system, we are led to consider a rather different range of hypotheses: The brain is changed importantly by experience with language; language is an instrument used by separate brains to exert biological influence on each other, creating through biological action at a distance a virtual brain distributed in the individual brains of all the participants in the culture; early experience with language affects cognitive operations that go beyond language.
Given the current state of linguistics, there are at least two predictable responses to the theory that language arises through parable. One is that it is trivially wrong; another is that it is trivially true, or at least obvious, something we have always known. I prefer the second view and am prepared not only to grant it but to make a gesture toward documenting it. Among literary critics and poets, it is an extremely common view, found in passages like the following from Fenollosa, where he proposes that a flash of lightning is a model (type) provided by nature of the grammatical sentence:
The type of sentence in nature is a flash of lightning. It passes between two terms, a cloud and the earth. No unit of natural process can be less than this. All natural processes are, in their units, as much as this. Light, heat, gravity, chemical affinity, human will, have this in common, that they redistribute force. Their unit of process can be represented as: term from which, transference of force, term to which. If we regard this transference as the conscious or unconscious act of an agent we can translate the diagram into: agent, act, object.
In typical fashion, Fenollosa sees grammar as arising by projection from reality rather than from story. But with only minor violence to such a passage, we can modify it into the claim that unmediated reality is not the basis of the projection; rather, abstract stories we use to make sense of reality are the basis of the projection.
Aristotle makes this claim explicitly in On Interpretation. Fenollosa views reality as objectively narrative, and so thinks reality is the basis of the projection to grammar. Aristotle, by contrast, says little about narrative structure but does understand that the projection begins not from reality but from conceptual structure. The form of language, Aristotle explains, arises by projection from conceptual structure.
In its broad outlines, this is not an uncommon view today. Neuroscientist Gerald Edelman, in elaborating the theory of neural Darwinism and neuronal group selection, offers a very brief proposal that grammatical structure arises from conceptual structure. Syntax arises from the projection of semantics onto phonology. Edelman rejects on neurobiological grounds any notion of a genetically programmed language module.
In its most simplistic version, the view that grammar arises by projection from story can be paraphrased colloquially by saying that sentences are small stories. The form of a sentence “is” the form of a story. Sentences are small stories that come from big stories. A noun is a person, place, or thing; a verb is an event; an adverb modifies an event; tense indicates temporal viewpoint and focus on a story; and so on. Such an ideational view of language is intuitive to anyone who has ever spoken, even though it cannot account for all of language.
At sophisticated levels, the theory that the origin of language is the projection of story is compatible with some other work in cognitive linguistics, in the minimal sense that it shares some basic assumptions with other work in the field. One of the most basic assumptions in cognitive linguistics is that linguistic structures can have conceptual bases. Perhaps the most immediately accessible analysis based on this view is Leonard Talmy’s analysis of ways in which force-dynamic conceptual structure is grammaticalized in English. A view of grammar as a dynamic system of interrelated grammatical constructions appears to be broadly compatible with the theory of grammar offered by Charles Fillmore and Paul Kay. The view that some basic grammatical constructions represent some basic human stories is compatible with work done by Adele Goldberg. Although Gilles Fauconnier and Ron Langacker do not speak exactly of the projection of story, they have in different ways analyzed with great insight and in great detail the ways in which projection from conceptual structure systematically underlies various aspects of grammar. Indeed, although Langacker and Fauconnier, like nearly everyone who works on tense, offer explanations of basic tenses that start from the temporal relation between the event reported and the moment of speaking, they nevertheless explicitly make the case for tense as the grammaticalization of viewpoint and focus. Various of Fauconnier and Langacker’s graduate students and colleagues have also done work on tense as arising essentially from focus and viewpoint.
It is often easy to assimilate the detailed work of a linguist who studies specific structures of language to the view that rudimentary grammar arose by parable. What is distinctive about my proposal is not the data or the analyses but rather the beginning frame. Thus, for example, although Comrie begins with a view of basic tenses as defined relative to the moment of speaking, a very little jiggle of his beginning assumptions can reposition his detailed analyses systematically to make his study compatible, I think, with the view that language arose through parable.
This impulse behind my proposal—make no unnecessary hypotheses—puts it at odds with the Chomskyan view that grammar arose because there evolved, with no help from natural selection and no help from preexisting human capacities, a species-specific, modular “organ” that is as specialized for grammar as the lungs, heart, and liver are for their particular tasks, expressed at the neurobiological level exclusively according to genetic instruction, and sharing nothing with other human capacities. There are alternative current varieties of this view, including Derek Bickerton’s “language bioprogram hypothesis” “that suggests that the infrastructure of language is specified at least as narrowly as Chomsky has claimed.” Bickerton draws his evidence for the “language bioprogram” from creole languages.
Any hypothesis that grammar is historically and individually the exclusive product of a special-purpose genetically provided mental organ for grammar rests upon a negative argument: We have not managed to explain how grammar arises through general processes; therefore we must posit the existence of a special genetic code that is entirely responsible for building in the brain a universal grammar organ, which will someday be located and understood in the way the retina and the lateral geniculate nucleus and primary visual cortex have been located and (partially) understood. The hypothesis of genetic instruction is not in principle impossible, but it is methodologically the hypothesis of last resort. It trades Occam’s razor for God’s magic hat: Against all odds, make the most cosmic and all-embracing extra hypothesis imaginable so as to solve everything at once—Let there be language.
Perhaps the main argument that grammar must arise in the individual human being exclusively from some special-purpose device, genetically coded and neurobiologically expressed, is that grammar is too arbitrary, subtle, and quirky to arise otherwise. But if the influence on language acquisition is not only the language an infant hears but also all of narrative imagining, including all of the systems from which narrative imagining recruits, there is plausibly an overabundance of sources for subtleties and quirks without conjecturing a special device to introduce them.
It is also argued that children obey subtleties of grammatical structure they have not heard, so that these subtleties must come from a special device. But if the influence on language acquisition is not only experience of language but also all of narrative imagining, including all of the systems from which narrative imagining recruits, then there is no poverty of relevant stimulus but rather a great wealth of relevant stimulus.
It is also argued that the subtleties and quirks of grammar are too universal to have arisen except through a universal mental organ for language acquisition. But story, projection, and parable are universal, and if grammar arises through parable, which recruits from universal systems such as vision, then there is no need to resort to a special conjectural mental organ to provide universality of structure.
Chomsky has been consistently unreceptive to the proposal that the hypothetical mental organ for grammar arose by natural selection. Stephen Pinker and Paul Bloom have argued in the most unhedged fashion that it does. Language, they assert, “is a topic like echolocation in bats or stereopsis in monkeys.” This extra hypothesis—that natural selection is entirely responsible for the hypothetical grammar organ—might seem to put Pinker and Bloom even further away than Chomsky from my hypothesis that language arose by projection of story. But not so. In their practical attempts to explain how natural selection could have produced a genetically coded mental organ, Pinker and Bloom implicitly embrace an account of language in which grammar begins from meaning. They write, “Language is a complex system of many parts, each tailored to mapping a characteristic kind of semantic or pragmatic function onto a characteristic kind of symbol sequence.”
“Mapping” is the critical term and concept in this assertion. It is usually the critical concept in any explanation of grammar as “encoding” something else, “signaling” something else, “mapping” certain structures, and so on. Yet the role of “mapping” in such explanations usually receives no comment, which is astounding. “Mapping,” which I have called “projection,” is a mental competence; it does not come for free in an explanation; it is instead the principal process to be explained. To speak of “mapping” is to make a theoretical commitment to a powerful and robust mental capacity of projection. Pinker and Bloom give various thumbnail sketches of such mappings that appear to me essentially compatible with a claim that grammar arises from the projection of narrative structure:
Noun phrases … are used to describe things. Similarly, a verb like hit is made into a verb phrase by marking it for tense and aspect and adding an object, thus enabling it to describe an event. In general, words encode abstract general categories and only by contributing to the structure of major phrasal categories can they describe particular things, events, states, locations, and properties.… Verb affixes signal the temporal distribution of the event that the verb refers to (aspect) and the time of the event (tense).
Pinker and Bloom’s explanation depends upon the existence of a mental capacity to project one kind of structure (story) onto something entirely different (vocal sound), thereby creating for vocal sound grammatical structure. Pinker and Bloom are assuming that the mental capacity for projection precedes grammar, or at least that grammar cannot arise without projection. Pinker and Bloom obscure the fundamental importance and complexity of this mental capacity by referring to it as simple “encoding” or “signaling.” This mental capacity—to encode, signal, map, project—is what principally needs explaining in an account of grammar. In explaining grammar, we are not discussing simple encoding such as “We call this dog ‘Harold,” where a particular entity is given an arbitrary label. We are instead discussing the projection of vast systems of narrative structure in such a way that complex categories of event correspond to grammatical categories of verb phrase. We are discussing the projection of systems of perceiving events with temporal focus and viewpoint to create systems of grammatical structure like tense. This is not a matter of giving a particular object an arbitrary label but of projecting structural categories to impart structural categories. Pinker and Bloom are thus assuming, as part of their explanatory machinery, the existence of a robust mental capacity to project one kind of thing onto another. In my view, they are right to do so. They are wrong, however, in assuming that this mental capacity is exclusive to language, rather than part of the mental capacity I call parable.
Pinker and Bloom imagine that the kind of conceptual information that is mapped is propositional structure. This is a misemphasis rather than an error. Story certainly involves propositional structure. The story structure of Mary throws the stone quickly carries the propositional information The throwing is quick. But basic grammatical constructions seem to come from basic stories, with agents and actions and objects and patients and viewpoint and focus and so on, as Pinker and Bloom seem implicitly to grant in their thumbnail sketches of the way grammar “encodes” conceptual structure. These basic grammatical constructions that arise from basic stories can secondarily be used for expressing lower-level propositional structure such as “Grass is green.”
Pinker and Bloom make it clear that they view their contribution as consisting entirely of the argument that natural selection explains the origin of grammar. They disavow any originality in their analyses of grammatical structures: “Any one of them could have been lifted out of the pages of linguistic textbooks.” But if we lay aside Pinker and Bloom’s argument about natural selection and look instead at their actual work in sketching how conceptual structure is projected to create grammatical structure, we see a treatment that appears to me (although almost certainly not to them) not far out of accord with the view that grammar arises from the projection of story.
The strong Chomskyan view of the origin of language asks us to believe that, against inconceivable odds, genetic instruction arose for a highly complex and sophisticated grammar organ, with no help from preexisting mental capacities and no help from natural selection. It asks us to believe that an extremely complex functional trait, language, is entirely genetic yet did not arise through the only mechanism of evolutionary genetics known to produce extremely complex functional traits, natural selection.
The astonishing unlikelihood of Chomsky’s model of the origin of grammar prompted Pinker and Bloom to argue that Chomsky is wrong. They embrace Chomsky’s picture in all respects except for their claim that natural selection is responsible for the origin of grammar.
Chomsky’s argument is weak because it asks us to accept an almost inconceivably unlikely event in the absence of any evidence for that event. Pinker and Bloom’s argument is weak in a different way: It skips briefly and vaguely over its central step. For natural selection to be responsible for the origin of grammar, we must have two events: First, some (minimal) genetic structure must arise that achieves penetrance to result in a trait of (minimal) grammar; second, this trait must occur in an environment in which it confers reproductive advantage. But that environment cannot be one in which a grammatical community already exists, since the origin of grammar is what we are trying to explain. The first event—the evolution of (minimal) genetic structure for (minimal) grammar—is not particularly hard to imagine, although there is as yet no compelling evidence of it. Let us look at the second step—conferring reproductive advantage.
Parable is a deeply basic capacity of human beings—we must grant this independently of any analysis of grammar. I have already sketched a scenario in which rudimentary grammar arose in a community through parable, not through genetic specialization for grammar. In that community, grammatical speech is a highly useful cognitive and cultural art. In that community, greater facility with grammar confers relative reproductive advantage, making genetic specialization for grammar adaptive. In my scenario, the origin of rudimentary grammar happens before any hypothetical genetic specialization for grammar. The grammatical expressions produced by the lone first genetically grammatical person are parsed at least in part as grammar by members of the community who have no special genetic instruction for doing so but who use parable to do so.
Do Pinker and Bloom offer a contrasting scenario, in which a lone first person with a genetic specialization for grammar has a reproductive advantage even though no one else in the community can recognize or parse any ofthe grammatical structure of her utterances? No. Such a scenario is not inconceivable—grammatical processing might assist the lone person’s memory, reasoning, or imagination, and so be adaptive indirectly. But Pinker and Bloom, whose analyses covertly suggest a view of grammar as rooted in meaning, adhere overtly to a school of thought committed to the strong view of grammar as autonomous of other cognitive processes. This stops Pinker and Bloom from offering a scenario in which reproductive advantage comes from the benefit of grammatical processing to other cognitive processes. It also stops them from offering a scenario in which members of the community without genetic instruction use parable (or any other cognitive processes) to recognize and parse grammatical structure in the utterances of the first lone genetically grammatical person.
Pinker and Bloom provide no alternative to the scenario in which the adaptiveness of genetic specialization for grammar depends upon the presence of a grammatical community. All of their speculations concerning reproductive advantage depend upon a community of speakers with rudimentary grammar. This is understandable. There is no obvious way in which a lone grammatical person would have a direct reproductive advantage in a community whose other members are completely incapable of recognizing or parsing grammar by any means, general or special. The utterances directed at her would not be grammatical. Her own grammar would have no audience, since none of the grammatical structure she produced could be recognized by companions. Of course, her grammatical utterance could still be understood as ungrammatical communication, and she could still attribute grammar to the ungrammatical communication directed at her, and there could surely be reproductive advantage conferred by communication; but it is difficult to see that there would be any additional advantage conferred by the grammatical component.
In a brief moment in their argument, Pinker and Bloom do consider explicidy the right scenario their argument needs: a grammatical community of kin. They write, “Geschwind, among others, has wondered how a hypothetical ‘beneficial’ grammatical mutation could really have benefited its possessor, given that such an individual would not have been understood by less evolved compatriots. One possible answer is that any such mutation is likely to be shared by individuals who are genetically related. Because much communication is among kin, a linguistic mutant will be understood by some relatives and the resulting enhancements in information sharing will benefit each one of them relative to others who are not related.” This is the only suggestion Pinker and Bloom offer of a community in which grammar is not widespread. It is still not a picture of reproductive advantage to the lone grammatical individual in a community whose other members are completely incapable of recognizing grammar by any means, presumably because Pinker and Bloom can see no such advantage.
But it is not true that such a mutation is likely to be shared by individuals who are genetically related. Consider the first genetically grammatical person. By definition, none of her ancestors or older siblings is genetically grammatical. If the genetic material expressed in her minimal grammatical competence arose by mutation from her parents’ genetic material, as in a copying error in making a sperm or egg, it is extraordinarily unlikely that her younger siblings would have that mutation. If it arose because error-free sexual recombination of her parents’ genetic material finally put together the right package, then later siblings would receive quite different genetic packages (especially if they do not have the same two parents). There is also the important difficulty that even if a later sibling had the right package, the first genetically grammatical infant would nonetheless still live in grammatical isolation (under Pinker and Bloom’s suggestion) during the period in infancy in which she develops grammar.
If evolution could think ahead, it would certainly see that producing a genetically grammatical community, even a small one, would be enormously useful to members of that community. Pinker and Bloom are certainly right about that. But evolution cannot think ahead. It cannot think even one step ahead. As George C. Williams puts it in Natural Selection, “Every step of the way, as Darwin made clear, had to be immediately useful to each individual possessor. No future usefulness is ever relevant.” A good natural selection story for the origin of rudimentary grammar must show that the first lone genetically grammatical human being had a reproductive advantage. If she was born into a community whose members had, by virtue of parable, a minimal capacity for grammar, then the reproductive advantage to her is obvious; but in that case, the origin of rudimentary grammar comes from parable, not from genetic specialization. Pinker and Bloom must disallow on principle the existence of such a community. They are therefore obliged to show benefit to the first genetically grammatical person, born into a community whose other members have zero capacity for grammar, who lack the ability even to recognize the existence of grammar, much less to parse it. This is the hard and all-important step in the argument. It is a step they skip over.
The least implausible picture Pinker and Bloom might have offered would be one in which the lone grammatical individual somehow develops grammar by using its grammar module in the absence of any grammatical utterances from other members of the community, grows up, passes the trait to one or more of her children, and converses with them in such a way that the minimal additional grammatical component that they add to communication confers, in some fashion, reproductive advantage to them, beyond the advantage conferred by the nongrammatical form of that communication, and skirting any costs in fitness associated with that genetic change, especially costs to the first lone grammatical person early in life, before she had a chance to reproduce and converse with offspring. This is a hazy picture, with central components missing, but nonetheless one in which genetic specialization for grammar could in principle have been adaptive even though rudimentary grammar was otherwise absent.
But there is no need to resort to such a picture. Very little grammatical competence would be needed to create an environment in which genetic specialization for grammar could be adaptive, and parable can supply a substantial grammar. The view that rudimentary grammar arises from parable provides an appropriate environment of evolutionary adaptiveness of the sort that any good natural selection story must have.
The hypothesis that human beings now living have some genetic specialization for grammar is an open question, to be decided on empirical grounds, principally in the study of genetics and neurobiology. The view of parable as the origin of grammar is not incompatible with this hypothesis. On the contrary, the parabolic account is the only one I see that straightforwardly provides the environment of adaptiveness that must be part of an adaptive account. In the parabolic environment, the lone genetically grammatical person would be surrounded by a grammatical community and would have an advantage over other members. Over time, genetic specialization in the species might take up some of the responsibility previously shouldered by parabolic mechanisms in the individual mind.
The view that parable explains the origin of language makes it possible to conceive of parabolic thinking and genetic specialization for grammar (if there is any) as historically connected in the evolution of the species and commensurate and reinforcing in the contemporary individual mind. Whether genetic specialization for grammar exists, to what degree it might exist, how it might be expressed in the individual brain under development, and how it might cooperate with other conceptual processes are all open empirical questions. But parable alone, without genetic specialization, gives us what we need for the origin of language.
The story I have offered reverses the view that language is built up from the sober to the exotic; that out of syntactic phrase structures, one builds up language; that out of language, one builds up narrative; that out of narrative, literary narrative is born as a special performance; and that out of literary narrative comes parable.
It works the other way around. With story, projection, and their powerful combination in parable, we have a cognitive basis from which language can originate. The dynamic processes of parable are basic to the construction of meaning and the construction of language. Story precedes grammar. Projection precedes grammar. Parable precedes grammar. Language follows from these mental capacities as a consequence; it is their complex product. Language is the child of the literary mind.
Parable is the root of the human mind—of thinking, knowing, acting, creating, and plausibly even of speaking.