Comments on Smith and Talmy
Read before The American Philosophy Association. Oakland, CA, March 29, 1997.
In a certain traditional and obvious sense, the human being is a substance, with observable properties and deeper-lying characteristics (properties and dispositions). That is, it has properties but is not a property, endures through time and space, and stands in causal relations.
Among its properties are the intentional ones: the flood of ‘ofnesses’ and ‘aboutnesses’ that qualify it on its way through the world. These directly qualify its experiences, the conscious events essentially interwoven into its life. They are ‘quasi-relational’ properties which are directly and distinctly identifiable in terms of what they are of or about, and are so in a way that does not rest on theories about them or their ‘objects’. (Of course any deeper understanding of them is another matter.) The most basic levels of human competence rest upon this pre-theoretic capacity to identify what our experiences are experiences of, and the first level of “phenomenological” work, as that has come to be understood in the Husserlian tradition, is description of objects of the main types of experience, and description of experiences of those types in terms of what they are ‘of’ and the manner of their ‘ofness’.
On this foundation, inquiry of a phenomenological type proceeds to other questions such as: What is it about our experiences themselves (their parts, properties, interrelationships and dispositions) that accounts for the fact that they have as objects the objects they do have? What are the differences ‘in’ experiences which account for the fact that the same object can be present to us in different ways? What is it for an ‘object’ to be, not merely an ‘object’, but an existing object, and how is existence of an object to be cognitively grasped. (See Ideas I, §135: “Object and Consciousness. Transition to the Phenomenology of Reason,” and the entire chapter immediately following it.)
Now the presentations by Barry Smith and by Leonard Talmy rest upon first level phenomenological work--whatever one may happen to call it, or not call it--and occupy themselves with questions of the next level up. The distinction between natural boundaries and fiat boundaries, as Barry discusses them, is a distinction between objects of our experiences. Representing Gibson’s view he says: “Each type of organism is tuned in its behavior to entities on a specific level of granularity within this complex hierarchy <the “world’>, to entities which together form what Gibson calls an ‘ecological niche’.... A niche embraces not only objects of different sorts, but also shapes, colors, textures, tendencies, boundaries (surfaces, edges), all of which are organized in such a way as to enjoy affordance-character for the animal in question. Thus the given features motivate the organism; they are such as to intrude upon its life, to stimulate the organism in a wide range of different though characteristically understandable and familiar ways.” (pp. 1-2; this type of view was also elaborated in M. Merleau-Ponty’s Structure of Behavior)
Human beings also have a niche, with “its basic organizing features...intrinsically comprehensible to the human organism...includ(ing) simple geometrical and topological relations, relations of identity, part and whole, as well as relations between qualities of different sorts.” (p. 2) They are “pre-tuned” in some measure to interact with these features of their environment, and they develop further such capacities through growing involvement with structures of their niche such as language and cultural forms. These structures further aid the person, not to create objects corresponding to its experiences, but to discern objects that could not otherwise be objects for them, some of which exist in reality while others do not. Among these objects are natural as well as fiat boundaries.
Leonard Talmy’s discussions are also concerned with space and spatial structures as objects of cognition or consciousness. They aim, in particular, at identification and description of human involvements with language, with lexical and syntactical entities of certain types, which produce or allow various actualities and possibilities of consciousness bearing upon space, motion, etc. as objects.
Regarding Barry Smith’s line of argument, I am in complete agreement that there are fiat boundaries (eg. the equator and the Utah state line). True, these would not exist “in the absence of all articulating activity on our part.” (p. 4) But our “articulating activity” is as much a part of the real world as are rivers and beaches, and often can do you a great deal more harm since they involve things like armies and all the paraphernalia of war. I would emphasize, perhaps more strongly than Barry does, that fiat boundaries do not exist by a mere act of consciousness or a private decision. Their esse is a good deal more than percipi, and the word “fiat,” though justifiable, has certain misleading connotations of immediacy and pure will. But perhaps he would not disagree.
I also would agree that our ontology should be extended to include fiat boundaries, such as the boundary of Utah. Such boundaries exist. I agree that “the standard distinctions which we can make between types of natural boundaries can be straightforwardly applied also to their fiat counterparts. (pp. 4-5) But I would say, somewhat more simple-mindedly, that the boundary of Utah has properties, eg. a certain length as well as a beginning and ending in time, and relationships to the laws and institutions, and the personnel and the physical equipment, of the state. There are standard methods of determining whether or not the boundary does have these properties. And anything that has properties exists. What more does one want? That they are not like something else in certain respects--eg. the natural boundaries of an island--simply has nothing to do with the case.
A similar point, too obvious to dwell on I hope, is to be made with reference to “fiat objects,” the things which have the boundaries, eg. the state of Utah. (p. 5) As one who loves neither deserts nor jungles, but a richly abundant land, I have never learned to regard “ontological parsimony” as a virtue. Of course, if something is already well understood and demonstrated to the satisfaction of all concerned, we should not introduce other entities to clarify or prove them. But that is almost never the case (I should just go ahead and say “never,” for indeed I believe it is so) in situations where ‘the principle of parsimony’ is invoked, and certainly it was not the case in the situation where Ockham brandished his ‘razor’ over Plato. To suppose that state and similar boundaries do not exist would certainly be thought, anywhere but in a philosophy classroom or convention, to leave a remarkable lot of things unclarified and undemonstrable.
When it comes to the section of Barry Smith’s paper on “Fiat Objects in Perception,” I am rather less comfortable. He says “the horizon is a component object of the visual field, and the latter may be defined, with Ewald Hering, ‘as the totality of real objects imaged at a given moment on the retina of the right or left eye’.” (pp. 8-9) I am afraid that such a statement as this can mean too many things. If the horizon is a real object imaged, not an image of an object, then I am not clear how it could also be described as a boundary “created by our acts of perception and by human cognitive processes of other sorts,” or as something created “by my very existence as a visually perceiving subject in a given location at a given time, as also by the parametric properties of my visual system, by topographical features of the location, and by the laws of optics.” (p. 8) I presume that two people can see the same horizon--though it will, in the manner of physical objects generally, appear somewhat differently to each--and that they can discuss and verify whether or not it has certain features. (Is the comet further above the horizon at this hour than it was last night?) I take it if one person walked away or ceased to exist the horizon--the same horizon--would still be there for the other. In short, it seems to me that the horizon is much more like a natural boundary than a fiat boundary, and thus determines natural events such as sunsets.
But I probably do not correctly understand what is being said. My suspicion that this is so is strengthened when I read that “the objects making up the visual field according to Hering’s definition <with Barry Smith’s agreement?> are primarily the surfaces of three-dimensional substances.” (p. 9) I would have to say that if the objects making up the visual field are to be real objects, as earlier indicated, then real objects are quite different than I have in the past thought them to be. And if, for example, we are still talking about horizons, among other things, I have to say I have never seen a two dimensional horizon nor, so far as I know, the surface of a horizon. So I am just quite lost here, and no doubt Barry can say something to straighten me out.
What Barry says about “Language-Generated Fiat Objects” and Leonard Talmy’s work thereon seems to me intriguing and important, and largely right, but I shall return to this topic shortly. Whether the structures that present themselves as a result of the “windowing” effects of language actually exist, as distinct from merely being present to the mind, and how, exactly, that “windowing” works (what are its component entities, relations and forces--what is“plasticity” and what, exactly, has it or exercises it), are of course the crucial questions. Barry Smith thus sees the issues clearly in these matters and raises the crucial questions. (p. 12)
So we arrive at his concluding words about “Truth,” which “has classically been understood in terms of a correspondence relation (i.e. of some sort of isomorphism) between a judgment or assertion on the one hand and a certain portion of reality on the other.” Barry notes that the “central difficulty standing in the way of this classical theory turned always on the fact that reality evidently does not come ready-parceled into judgment-shaped portions of the sort that would be predisposed to stand in relations of correspondence of the suggested sort.” (p. 12) Can we be helped by looking at the “language-induced fiat boundaries” studied by linguists, and then treating “judgment itself as a sui generis variety of drawing fiat boundaries around entities in reality” (p. 13) in such a way as to get correspondent truth makers for our judgments? It seems, to Barry, that we can. We can replace the confused notion of ‘conceptual reality’<reality as conceived, as present?> operative in cognitive linguistics with “the geographer’s notion of reality subject to fiat articulations.” This will mean, I take it, that the articulations projected from linguistic structure are actually there to serve as truth makers for sentences in natural languages, and that we are “in a position to exploit” the resources of cognitive linguistics “to produce a truly adequate account of truth for natural language in correspondence-theoretic terms.” (p. 13)
But can we replace the confused notion--confused about exactly what we have in it--of ‘conceptual reality’ with the fiat articulations of the geographer? I would have to be very hesitant about it. The two seem to have a lot of dissimilarities. The ‘conceptual reality’ of the spatial layouts and orientations “windowed” by language use do seem to be much more in the category of objects whose esse is percipi. We have an array of considerations about human behavior in the real world that give a substance to the geographer’s boundaries with no parallel in the spatial forms present to linguistically framed cognition. The specific boundaries of the geographer are, accordingly, intersubjective in a way that the spatial forms are not. This is not a conclusive objection, in my opinion, but it is troubling for the prospects of such a transfer as Barry Smith suggests.
Moreover, I cannot agree with Daubert--if this is his view--that “in the absence of the judging activity an entity of the given sort <the truth maker?> would in no way be demarcated from its surroundings” nor would it have the internal differentiation required for it to play its truth making role. True, when we explicitly form an empirical judgment about something already before us, the ‘something’ does then present to us a structure which we are not aware of before we form the judgment. For example, when I judge, looking, that my scanner is by my computer on my desk, these various objects stand out to me in the relationship indicated. But the idea that they enter those relations that make my judgment true in virtue of my lookingly judging them to be in that relation is one I find hard to accept--or even understand. Being in those relations just doesn’t seem to me to be a fact, being what it is, that could be brought about by a perceptual judgment, being what it is. Yet it is there, and we can see it over against our judgment about it if we wish to, as we do constantly confirm and disconfirm our beliefs by comparing them to what they are about. I realize that many of our brightest philosophical lights insist, for what they regard as good reasons, that you cannot do such a thing. But we do it all the time. To speak of this as involving a “God’s eye view” is hardly helpful, and to make one’s epistemology dependent upon the non-existence of God.
With this we turn to the work of Leonard Talmy:
To begin with, his research project clearly depends, if I understand it, upon what I have called the “first level of phenomenological work.” He would not be able to get started without first identifying experiences in terms of their peculiar types of objects (spatial forms, etc.), on the one hand, and the peculiar involvements of those same experiences with language and specific types of linguistic forms, on the other. Further, one must be able to identify dependencies of the objective bearings of the experiences upon the linguistic forms involved. Within the dependencies the crucial feature of plasticity emerges, as marking a distinctive range of potentialities as to what objective (spatial) structures can and cannot be ‘represented’ in experiences involving distinctive ranges of linguistic forms.
Now I do not object to this at all, but find it to be completely appropriate--and indispensable. I only object when this is done and not owned up to, which it seems to me is not the case in Leonard Talmy. He seems to be quite forthright about it. But not everyone is. In my opinion “doing it without owning up to it” seems to me to have been standard practice for decades among those philosophers who thought of themselves as analyzing meaning or language. They were totally dependent upon first level phenomenological work, but their epistemology could not allow for it.
Now on the basis of first level phenomenological work of the sort indicated, Leonard proceeds to draw and theoretically organize basic linguistic distinctions--such as that between the “Open-Class Forms” and “Closed-Class Forms” and the interplay between them--and to apply them to Topology: to space and motion as objects of certain linguistically framed experiences. His theoretical organization of correlations between linguistic forms, spatial structures as present, and the experiences within which the two are correlated, seems to me, as a non-expert with reference to the field of Linguistics, to be highly illuminating and well-founded. I regret I am not in a position to say anything more on this point. I agree that our experience of language somehow directs the flow of intentionalities (ofnesses and aboutnesses) toward our world. This is a vital area of research, regardless of what is to be made of it philosophically, and the linguist is the one to research the regularities and formulate their laws.
With respect to philosophically ‘deeper’ issues, I have a number of concerns:
First, I am worried about treating language as a “cognitive system,” perhaps on a par with other “cognitive systems” such as “visual perception” or “reasoning.” Language understood as some system of linguistic forms doesn’t seem to me to cognize anything in the way perception and reasoning, for example, do. Of course it all depends on how we come to understand “language” in some manner deeper than may be required to do fine work in linguistic theory, and that is a question that involves the deepest issues in ontology and the philosophy of mind.
Second, I presume that we must cognitively experience language before we can experience other things by means of language. What is fundamental, therefore, in the theory of cognition, will not be experience by means of language. It is extremely unlikely, in my opinion, that research with respect to it will teach us much about “the general character of conceptual structure in human cognition,” as Leonard elsewhere says, because it is precisely that general character that would have to be understood before we can understand how cognition of language itself is possible and how cognition of language functions in consciousness of things other than language. I think that cognition of language cannot itself be made possible by the cognition of language, and we need to clarify what cognition of language forms is before we can understand how that cognition makes possible the cognition of other things. “The cognitive representation evoked by the sentence” points back to the cognition of the sentence. An alternative might be that the “evoking by the sentence” does not involve cognition of the sentence, but it seems to me it clearly does, and references to hearing the sentence or seeing the sentence in cognition of, eg. space, surely bear this out. I realize that the main traditions in recent philosophy (Wittgenstein, Ryle, Sellars) have held that knowledge of language falls into some special category, where ‘knowledge’ of it is to be excepted from questions we might ask of knowledge generally; but I am unconvinced, and I think that the natural descriptions given by the linguist indicate that quite straightforward cognitive experiences of language are presupposed by representation of other things by means of it.
Third, it seems to me that the organizing tendencies that result in the appearances of spatial forms and other objectivities corresponding to linguistic forms are best understood as falling on the side of our experience of language (in its environing world) and not on the side of linguistic forms as objectively identifiable. In his paper on “The Relation of Grammar to Cognition” Leonard Talmy states: “The terms ‘grammatical’ and ‘lexical’ as employed here require some immediate elaboration. The distinction between the two is made formally--i.e., without reference to meaning--on the basis of the traditional linguistic distinction between ‘open-class’ and ‘closed-class’. A class of morphemes is considered open if it is quite large and readily augmentable relative to other classes. A class is considered closed if it is relatively small and fixed in membership.” (p. 166) In his presentation he speaks of Open-Class items as “unconstrained as to what they can refer to,” and Closed-Class items as “highly constrained.” Now to put my concern simply: It seems to me that there is no way of treating such “constraint” with regard to reference as a “formal” feature in the sense he indicates.
And of course the same point must be made with reference to “plasticity”: “Plasticity is a design feature of language because it allows a limited set of symbols to refer to a vastly larger domain of conceptual material. Closed-class forms with abstracted schematic--including topological--referents are a design feature of language because they structure its conceptual material. The schematic structuring system of language seems to have similarities to the structuring system of other cognitive systems such as visual perception. Mathematical topology may well be a theoretization of topology-like structuring already built into the fabric of language and of other cognitive systems.” (Conclusion of his presentation)
The idea of “structuring the conceptual material of language” is then, in turn, dependent upon the “constraints” and the “plasticity” ‘in’ language. And the questions as to whether the “structuring” has the effect of producing the (spatial and other) properties of objectivities, or only making them present to us--Do we get from language the representations, and is that all we get, or reality revealed with the representations? etc.--follow, in part at least, from what seems to me to be the basic obscurity of how the “morphemes” are supposed to function. (Thus statements like:“One main characteristic of language’s spatial system is that it imposes a fixed form of structure on virtually every spatial scene” (II.3. of Leonard Talmy, “How Language Structures Space”) remains fundamentally obscure as to exactly what “imposed” “scene” etc. mean.)
Husserl located the “action” on the side of the mind which experiences the “morphemes,” and with his accounts of “Significations,” “Motivation” and so forth (see Ist, IVth and Vth of the “Logical Investigations” and elsewhere), he tries to provide a clear and literal account of what “constraint,” “plasticity,” and “structuring” might mean. But then he does not make language essential to cognition as such. Given an appropriate distancing from cognition, it seems possible to return language to a realistic role in cognition by, among other things, providing an account of cognition of language itself. That he tried to do.