Reflections on the Evolution of Human Language:
4. On the Evolution of Subject Matter
Introduction
The earliest content
Situations
The outside world
The "saying mode"
Superindividualism
On "knowledge of language"
In the "Thoughts on reading Heryanto" sequence (Grace 2007) one question that kept recurring concerned what I called
the "prototypical function" of language. By this, I meant what language is ultimately designed for, and this
seemed to boil down pretty much to what kinds of messages it was designed to convey. I acknowledged that I'd
always simply taken it for granted that language was essentially an instrument for encoding and decoding factual
information (and what I mainly had in mind was the kind of information normally represented in the form of
propositions). However, with more careful consideration there seemed to be no basis for presupposing anything of this sort.
Of course, that means that our efforts to understand the human brain or cognition or what it is to "know" a language
at the very least shouldn't start from any presupposition that that's what language was intended for. I think of the
question of the nature of language as coming down ultimately to a question about its evolution: about what
the brain--or at least its language-enabling parts--was designed to do. What capabilities do humans possess
that result entirely or in part from our adaptation for language? What do our brains enable us to do
that we would not be able to do but for the process of language evolution? In other words, what exactly
does humans' innate language capacity consist of?
If we could get sufficiently specific answers to that question, they would provide a helpful guide in a whole
array of research fields having to do with different aspects of human cognitive functioning.
What I offer here are some thoughts and speculations about how language is designed to present, and represent, subject matter.
REFERENCE
Grace, George W. 2007. Thoughts on reading Heryanto. Internet WWW page at http://www2.hawaii.edu/~grace/Heryanto.html.
It would seem the first question to ask would be where it all began.
It seems important to remember that when our ancestors began this evolution
they and their closest kin were nearly identical. The only difference that's
really significant for us is that there were behaviors that were advantageous
for our ancestors and either lacking in their kin or not equally advantageous
to them. Thus, there was a common starting point and then something intervened
that led our line to diverge.
What was this common starting point? This question could be considered from many
angles of course, but our concern here is what they communicated about. On this,
the ideas of Gregory Bateson provide (what is for me, at least) a stimulating new perspective.
Bateson had an ability to step back and provide fresh and insightful overviews of a variety
of subjects with which he had experience. He had extensive ethnographic field experience and
research on mental disorders such as schizophrenia, but what are of interest here are his
studies of communication in a number of different mammal species. Some of his general
observations about mammals provide what seems to me to be a good place to start. I have
in mind particularly that presented in the following quotation:
In general, the discourse of animals is concerned with relationship either between self and other or self and environment.
(Bateson 1972: 141)
This suggests the hypothesis, then, that at the time when the human ancestors began their separate evolution,
their communication consisted of behaviors that conveyed this kind of information. But at some point at least
some of these behaviors came to offer new advantages, or else the kinds of advantages they offered became more valuable.
The evolution can presumably be thought of(1) as a matter of the selection in each generation of advantageous
behaviors. What I mean is that in each generation there were behaviors of some kind that proved to be advantageous
to the particular individuals in that population who had the requisite ability (or inclination). "Advantageous"
in this context, of course, means advantageous to their genes--that these individuals had more than their share
of descendants so that the genes that found expression in the advantageous trait were found in an ever-increasing
proportion of the population.
Of course, not all of the advantageous behaviors led in the direction of language, but some did. That is,
some were antecedents of the behaviors that constitute use of language today.
What kinds of advantages might have been involved? What kinds of advantages might have been enjoyed
by those individuals who were particularly effective in communicating about relationships?
My guess here (based in large part on my reading of Dunbar and Malinowski)(2) is that it reinforced
interpersonal connections and even multiplex group connections. That is, it helped to create rapport
and a sense of solidarity. In so doing, it would have enhanced group cohesion and probably made more
cooperative activities possible. And of course as time has gone on, it has become possible to coordinate
increasingly complex activities of increasingly large numbers of individuals.
The hypothesis, then, would be that those who performed the same kinds of communicative behaviors more
frequently or with greater skill enjoyed a selective advantage. But why would that have been true in our
ancestral population and not the populations of their most immediate primate kin? The most likely answer
that I'm aware of is that provided by Dunbar--that our ancestors were adapting to a changed habitat with
larger groups and greater dispersion so that communicative behaviors that could serve in lieu of face-to-face
interaction acquired new value.(3)
Of course, what existed then in the way of antecedents to language was primitive in the extreme.
A lot had to change before we were anywhere near language as we know it. But still, I find it probable
that this original content--information about relationships--has remained the basis upon which today's
superstructure was built--and upon which it continues to rest.
NOTES
1. At least as far as natural selection is concerned--I'm ignoring the possible role of sexual selection (but see Miller 2000)
2. I've depended particularly on Dunbar 1972 and Malinowski 1923/1946 and 1935/1965. See also Grace 2003.
3. Note that this must have meant a greater dependence on audible components, and therefore a particular selective advantage for them.
REFERENCES
Bateson, Gregory. 1972. Steps to an ecology of mind. New York: Ballantine Books
Dunbar, Robin. 1996. Grooming, gossip, and the evolution of language. Cambridge MA: Harvard University Press.
Grace, George W. 2003. Robin Dunbar's Social Bonding Hypothesis. Reflections on the evolution of human language, number 1. Internet WWW page at http://www2.hawaii.edu/~grace/dunbar.html.
Malinowski, Bronislaw. 1923/1946. The problem of meaning in primitive languages. Supplement to C. K. Ogden and I. A. Richards, The meaning of meaning. 8th edition. New York: Harcourt, Brace, and Company.
Malinowski, Bronislaw 1935/1965. Coral Gardens and their Magic. Vol. 2, The language of magic and gardening. Bloomington: Indiana University Press. (Originally published in 1935 by George Allen & Unwin, Ltd.).
Miller, Geoffrey F. 2000. The mating mind: How sexual choice shaped the evolution of human nature. New York etc. Doubleday.
What first got me thinking about the way situations are involved in speech acts was a problem relating to translation.
At that time I was trying to understand what translation was--ideally, to come up with a general definition. The term "translation"
seemed always to refer to some kind of equivalence between two things (usually texts or utterances). In fact, it was often casually
assumed that the equivalence was one of meaning, but that often was clearly not the case (consider the "translations" of the "meaningless"
verse, Jabberwocky). Anyway, I was having trouble figuring out just what all the kinds of equivalence actually were and what they had in common. I've
discussed all of this more thoroughly in Grace 1987 (and also Grace 1989), but I'll give a brief review here.
- I started from what I took to be the common conception of translation--of the prototypical act of translation--which
was approximately the following:
- 1. Translation is prototypically a relation between texts (by "text" I mean a particular wording--not necessarily written).
The relata consist of a source-language text (as input) and a target-language text (as output).
2. The relation involved is understood to be one of equivalence of "meaning".
3. In the prototypical case, equivalence of meaning can be determined by comparing meanings of the two
texts arrived at by computing the meanings of the constituent lexical items and the meaning(s) of the
construction(s) in which they appear.
But then I encountered the problem of which I speak--that speakers often found the resultant translations unsatisfactory
because they felt that they didn't correspond to what target-language speakers would actually say in
the particular situation. This complaint seems to be especially frequent when the texts are not expository
prose documents, but informal conversational speech.
Anyway, I was confronted by the fact that in many cases translators found it preferable to reject the
computed meaning equivalent in favor of an expression that was felt to be what target-language speakers
would say "in the same situation". In fact, it seemed that as a general rule, this what-they-would-say-in-the-same-situation
principle was perceived as taking precedence over the computed meaning principle.
But if what-they-would-say-in-the-same-situation was the overriding concern for translation, it seemed
important to give it some further explication. How can we know what they would say? And above all,
what is the same situation? In fact given any speech act, how do we go about specifying the "situation"
to which it is the response--which presumably is to be understood as the culturally-appropriate response?
To illustrate the problem, consider a seemingly simple example. Suppose we hear a man say, "It's raining",
and are asked what situation evoked this response. The first impulse might be to answer that it's the rain itself,
or possibly the fact of his perceiving the rain.
But then suppose we see another man looking out a window at a landscape on which a heavy rain is falling.
And he says nothing at all, or makes a remark about last night's dinner, or any of an infinite number of
other possible things. In what way was he not in the same situation as the first man: what factor was
present in the first case and lacking in the second? Clearly, it is a matter of intention--the first man
wanted to tell someone (the addressee) about the rain whereas the second had no such desire. To deal with
this complication I proposed distinguishing the objective fact that it was raining as the "objective situation"
and the subjective state that led to his decision to tell whoever it was about the rain as his "subjective situation".
In short, I decided that the situation that would have caused our hypothetical man to say "It's raining" was
in fact a subjective situation. Of course there are many different circumstances in which those words
might have been used, but what we might think of as the most straightforward--or default--case can serve
as an illustration. In this case the speaker's subjective situation would involve possession of the
information that it was raining and some feeling toward the addressee that prompted him to share that
information with her (to keep the pronouns manageable, let's make the addressee a woman). To imagine
what speakers of another language would say "in the same situation", we must imagine them in a comparable
state of mind (i.e. "subjective situation"). Furthermore, this is universally true; the situation
that evokes any speech act in any circumstances is the subjective situation.
Why do I bring the business of situations up in a sequence about message content? My point is to show the continuity
between subject matter in today's language and that envisaged for the earliest stages. I proposed
in the conclusion of the preceding segment that the original subject matter--"information about relationships--has remained the
basis upon which today's superstructure was built--and upon which it continues to rest". One point is that the
speech act itself, the context in which it occurs, and the way in which it is executed combine to constitute a
message about relationship. In our default interpretation of the "It's raining" speech act, we might describe
this message as something like: "I feel my relationship to you to be such that I desire to be helpful to you,
and being conscious that you would want to know that it's raining, I am informing you of that fact." The feeling
thus described is what the speaker is representing to be his subjective situation.
A lot has obviously changed between the communication of the earliest stages and language as we know it today.
For one thing, the separation between what a speaker represents his/her subjective situation to be and the subjective situation
that actually evoked the speech act has widened. There has surely been an increased awareness by the participants
that what was happening between them was communication accompanied by an increased appreciation of the potentials it offered.
By saying
it was communication, I mean that one individual's expressions of feelings (of his/her subjective situation) were evoking
responses from others. And this would inevitably have led to the further recognition that a speaker had the possibility of
manipulating the behavior of others by deliberately producing artfully-chosen expressions. In other words, one could represent
feelings different from what one was actually experiencing. By the same token, there was the realization that others might
be doing the same thing with their signals. At any rate, there must have been an increased awareness of the fact of
communication, and consequently of some of its potential.
Thus a difference has grown between the effective subjective situation--the subjective situation that
actually underlies the speech act--and the ostensible one--what the speaker is trying to make it appear to be. However, it
is fair to say that no hearer can be fully satisfied of having understood a speech act unless s/he has some
sense of having understood the speaker's motive--his/her effective subjective situation. In many cases this
motive seems fairly obvious; then it is likely to be taken for granted without further thought. But in some
cases, of course, it is much puzzled over.
The "same situation" of what-they-would-say-in-the-same-situation translation is, of course, the
effective subjective situation. The translator must infer the situation of the source-language speaker
and resort to some kind of knowledge of associations between subjective situations and linguistic expressions
(or speech acts) that is available to target-language speakers. The whole question of how speakers learn
which linguistic expressions (or speech acts) are to be associated with which subjective situations is
a question worthy of interest.
A second major kind of change that I want to bring up is the enormous expansion of the realm of
objective situations. The awareness of the possibilities offered by the freedom to design one's
messages would probably also have drawn more attention to the distinction between the subject
matter of the utterance on the one hand and the intended audience (or eventually, addressee) on the other.
In the earliest stages, there seems to have been no way to talk to a about b.
As far as the inclusion of subject matter beyond feelings toward an addressee is concerned,
that seems always to have been occurring in a limited way. Individuals had always expressed feelings
toward things other than an individual with which/whom they were immediately interacting: weather phenomena,
external threats (e.g. the famous vervet responses to predators).
A third innovation--possibly the most intriguing of all--is what might be described as a new communicative
mode that has been superimposed on the regular relationship message. The mode still used for representing
one's subjective situation might be thought of as communicating by suggestion, as hints, clues; it might be
described as acting out the message rather than saying it. This is a
direct continuation from the postulated earliest stage as we hypothesize it to have been.
The new mode is more straightforward and explicit--it is used in what we normally think of as actually saying a thing.
I've heard this new superimposed message referred to as the "text" and the accompanying
relationship message as "subtext". It may be of significance that this new mode is most closely
associated with communication concerning objective situations, although it can certainly be applied to subjective
situations as well.
One final note: Written language has assumed such a prominent place in the modern world that people today
often find it hard to think about language at all without thinking about its written form. It's no doubt
the case that written language makes the relationship message (the subtext) more obscure, especially when
the writer has only a very vague and general conception of the nature of the eventual readers. But I think
we must acknowledge that it is still there. Even the writer of a book has an attitude toward the prospective
readers: some conception of what kind of people s/he expects or wants to attract as readers and what kind of
impression s/he wants them to receive of the subjective situation that underlay the book. And I believe that
readers generally feel some need for a sense of "where the writer is coming from". Thus, it seems fair to say
that there's always some kind of relationship message at base of the communication. But in any case
mass-produced writings were unimaginable during
the period when the brain was evolving its capacity for language.
REFERENCES
Grace, George W. 1987. "What they would say in the same situation". Ethnolinguistic Notes, Series 3, no. 31, Printout)
Also (1996) Internet WWW page at http://www2.hawaii.edu/~grace/eln31.html.
Grace, George W. 1989. The association of situations with linguistic expressions. Ethnolinguistic Notes, Series 3,
no. 35, Printout) Also (1997) Internet WWW page at http://www2.hawaii.edu/~grace/eln35.html.
This segment posted 8 April 2008
In the preceding segment I noted that one of the most critical developments in the evolution of subject matter has
been an enormous expansion in the ability to refer (or allude) to things in the outside world--that is, to things
external to the speaker, or as we've been describing it, to his/her objective situation. The aim of this
segment is to provide some ideas about how that might have come about through gradual steps--that is, without the
intervention of anything bordering on the miraculous.
Since this talk of "subjective" and "objective situations" is unaccustomed, it might be worth a brief
terminological review. A person's "subjective situation" is what is going on inside him/her while his/her "objective
situation" is what is outside--and therefore potentially accessible to others. It should be emphasized that neither
the entire subjective nor objective situation is actually in play in any particular speech act. In the preceding
segment, I proposed a distinction between (1) the effective subjective situation--the particular aspects of
the speaker's overall state that motivate her/him to execute that speech act--and (2) the ostensible subjective
situation--what the speaker is trying to make the effective subjective situation appear to be. We might also refer
to an effective objective situation: whatever part of the overall external environment the speaker intends to
include in the subject matter of the utterance. (If it turned out to be useful, we could conceive of an ostensible
objective situation as what's involved when the speaker misrepresents something in the objective situation).
It would be hard to exaggerate the consequences of these expanded abilities. The subject matter of our speech
today routinely includes things other than our own feelings, even things toward which we have no discernable
feelings at all. In fact, aspects of speakers' objective situations play so familiar a part in our language use
that we probably tend to assume them to be the prototypical subject matter. That is, it seems easy to assume that
talking about such objective situations is exactly what language was for in the first place.
However that, of course, isn't the case: it has taken fundamental changes to bring us to this point, and it
isn't easy to imagine the steps that brought them about. What I aim to do here is simply suggest a few innovations
along the way that probably helped set the stage.
To begin with, it seems likely that one step forward involved an enhanced awareness of the possibilities of
communication. This might have even had to start with nothing more than an awareness that one's expressions
of feelings were in fact communicating--that is, that others were perceiving and responding to them. At any
rate, at some point some individuals must have become more conscious of the fact that their speech acts (I'll
use the term "speech act" for whatever communicative acts occurred at the time) were eliciting responses from
others and, furthermore, that speech acts containing different messages elicited different responses. Some must
then have realized that it was possible consciously to manipulate their own communicative behavior and that in
doing so, they could elicit different responses. It's easy to imagine that the realization that they had some
freedom deliberately to choose their own messages would have tempted some individuals to experiment with the
potential for novel messages.
A particular ability that would presumably have had to be acquired for this tactic to be of more than very
limited use was simply that of communicating specifically to a about b, where b was
another individual or thing--that is, communicating to an intended addressee about a subject matter unrelated
to him/her. Before this step, subject matter seems to have been restricted to expressions of one's feelings
toward an individual or individuals with whom one was interacting.
There does appear to have been one limited exception: the expression of feelings such as anger, fear, or
frustration toward things such as external threats or weather phenomena. However, there seems to have been
no way to communicate the latter feelings specifically to some particular individual or individuals. In sum
then, the ability to address a message to a particular addressee when that message was about something other
than the speaker's feelings toward that addressee must have been an important innovation.
What I'm proposing then is that (1) the ability to address messages specifically to a particular addressee,
(2) the ability to extend the potential subject matter of these messages to things other than feelings toward
the addressee, and (3) the awareness that messages could be deliberately manipulated to elicit different responses
all encouraged an exploration of the possibilities of different messages and experimentation with novel ones.
Possibly the experimentation was initially directed simply at misleading the addressee as to the speaker's feelings
toward him/her or motive for speaking at all. However, it's easy to imagine how experiments might fairly soon have
gone further and discovered the possibility of implicating external conditions in the motivation. For example, one
can imagine an individual calling attention to some external threat where this act would simultaneously suggest
awareness of the threat and desire to make the addressee aware of it.
Of course, I don't intend to suggest that any message at this stage could have been at all specific--just something
to warn the addressee to pay attention. But still the threat is part of the speaker's objective situation, and its
presence is implicitly part of the message.
However a very simple message--in this case roughly "there's something that has me frightened that I want to call
attention to"--could have proved useful in multiple ways. In fact, nothing more than the ability to call attention
to particular environmental features might prove useful in developing cooperative undertakings--in eliciting the
cooperation of individuals who might otherwise have remained indifferent to the ongoing activity.
If the use of such messages produced more effective cooperation, it should have paved the way to more varied and more
effective techniques in such basic activities as hunting and gathering food. As a consequence, the affected communities
should have experienced advantages such as more generous food supplies. And this would presumably have provided a selective
advantage to those individuals who displayed the greatest skills in employing the messages, and led in subsequent generations
to increased levels of skill and frequency of their use.
But although so highly general a message could have served as a starting point, the possibilities for further advance
would have been very strictly limited until there was an increase in the repertoire of possible messages. This increase
would presumably have involved message differentiation. By this I mean that such a simple message--doing nothing more
than inviting the hearers' attention--could have opened doors in multiple directions. For example, we might expect
differentiations indicating something about where the attention should be directed or who was being addressed. Of course,
the first steps would surely have been very small ones: most likely, I would think, messages manifesting some form of
deictic or possibly pronominal coloring.(1) (And of course we might also expect more subtle
distinctions of speaker affect beyond just fright versus its absence).
Anyway, such developments would have expanded the spectrum of what we've been calling the "ostensible objective situations"
that the communications apparatus is capable of distinguishing. But more to the point perhaps, note that this would have
been an expansion in what we may think of as the repertoire of possible subjective situations. What the speaker's
utterance would have conveyed was that s/he had (i.e., was in the situation of experiencing) some feeling in which
this particular external situation was involved in a relevant way.
I should probably make it clear that I'm not assuming any kind of analyzable utterances at this point. What I am imagining
is that new utterances (still not productively analyzable) were sometimes added to what was essentially a fixed inventory.
There must have been a gradual increase in ways in which such additions were formed, but surely all would have involved
modification of existing members of the inventory (see the discussion of possible processes in Grace 2008).
In Grace 1987 and Grace 2005a I've discussed the role of language in constructing and representing our knowledge of
the world around us--of our encompassing reality. To be more accurate, what our language provides is really a model--rather
than an exact representation--of reality, but one that greatly enhances the ability of our species to function effectively
in the world. At the beginning, this might have involved nothing more than calling attention to certain phenomena or certain
actions and thereby giving them comparable saliency in the realities of both speaker and audience. This could have led to
shared awareness of particular facets of the external world and the sharing of certain skills.
I don't have much to propose about the nature of these "facets" of reality. One thing I would suggest is that we not
automatically think in terms of nouns. I would expect them to be more reminiscent of verbs and maybe adjectives.(2)
At any rate, what I'm proposing is a gradually expanding ability to share perceptions. I'd expect that most of these
perceptions would have represented situations or tasks in the immediate environment, but that shared conceptions of a
more permanent reality would have gradually emerged as a by-product. And of course I propose that the ability to share
such perceptions would have been advantageous and consequently favored by natural selection.
Very gradually at first as I picture it, but with accelerating speed thereafter, an individual could call others' attention
to entities and structures that s/he was distinguishing in the external environment. In this way, individuals' models of
the world would have come to be shared, and augmented by the insights of others, and a commonly perceived reality would
have gradually taken shape. And this reality may have expanded increasingly fast as the possibilities of metaphor and
metonymy, for example, came to be appreciated.
There's one further point to bring in here. I've also proposed (cf. Grace 1987 and Grace 2005a) that these communications
probably clustered around particular practical interests, for example, around basic economic activities and the knowledge
and skills that advanced them. The point is that the reality construction should not be imagined to be a philosophical
undertaking directed at general knowledge, but rather to have been focused on areas of special practical interest. This
would mean that the antecedents of what I've called ways of talking about things and Andrew Pawley (1991) has
called subject matter codes were distinguishable from a very early stage. But I've discussed this elsewhere--quite
recently in Grace 2005a--so I won't attempt to enter into more specifics here.
To sum up: The purpose of this segment has been to offer some suggestions about how the subject matter of messages that
consisted at the outset simply of an expression of the speaker's feelings might have gradually expanded to include more
and more information about his/her objective situations. It has been concerned with subject matter rather than the apparatus
available for expressing it, but it's probably worth mentioning here that before anything more than very simple objective
situations could be represented, there must have been some progress toward analytic processing and composition. I suggested
some of the possible steps and factors in this process in the series on "The emergence of analytic processing". For a quick
summary, see Grace 2005b.
I should also point out that the steps proposed here still wouldn't bring the evolving language beyond the old mode of
communicating by hinting at things rather than by explicitly saying them. In other words, the message might be thought
of as only the depiction of a kind of conceptual situation or event that leaves the hearer to figure out its relevance.
The expansion from this "hinting" mode to the kind of "saying" mode with which language users today are familiar is a
separate topic that would take this segment too far afield. However, it shouldn't be imagined that they didn't overlap
in time. No doubt some innovations leading in the direction of "saying" had probably occurred before the constructed
realities (the kinds of conceptual situations or events that could be expressed) had become in the least elaborate.(3)
NOTES
1. It also seems likely that some level of awareness of likenesses between different situations may have led to a kind of
metaphorical extension, resulting perhaps in homonymy. There might also have been some exploitation of natural associations
in a sort of metonymy.
2. In fact, I've always suspected that the centrality that seems to be accorded to nouns by
MSL-speakers (4)
is probably a culture-centrism. MSLs seem to assign a particularly prominent role to nouns and nominalizations.
3. As I was contemplating what I'd written here, it occurred to me that it could be seen as connecting in some ways
to the model in Karl Bühler's 1934 Sprachtheorie (Bühler 1965). There (to paraphrase rather liberally) he
described a triad consisting of speaker, addressee, and the real world, with linguistic utterances having functions
with respect to each. There are correspondingly three functions for which language is used: (1) the expressive
function (to express his/her own feelings) (2) the Appell function (maintaining contact with and persuading addressees)
and (3) the Darstellung (representational) function (conveying factual information or the like, where "the like" here
is intended to cover a multitude things including lies, questions, etc.)
In the view presented here, the expressive function was present from the beginning, at least the beginnings of the Appell
have been suggested, and the representational function corresponds to our "objective situation".
4. For explanations of what is meant by "Modern standard language" (MSL), see msl1
or msl2
REFERENCES
Bühler, Karl. 1965 (1934). Sprachtheorie: Die Darstellungsfunktion der Sprache. 2., unveränderte Auflage. Stuttgart: Gustav Fischer Verlag.
Grace, George W. 1987. The linguistic construction of reality. London: Croom Helm.
Grace, George W. 2005a. Reflections on the evolution of human language: 3. A "Swiss-Army-Knife" Conception of Early
Language Evolution. Internet WWW page at http://www2.hawaii.edu/~grace/knife.html.
Grace, George W. 2005b. Reflections on the evolution of human language: 2. The Emergence of Analytic Processing:
4. Summary of suggestions toward a hypothesis. Internet WWW page at http://www2.hawaii.edu/~grace/sum.html.
Grace, George W. 2008. Reflections on the evolution of human language: 2. The Emergence of Analytic Processing.
Internet WWW page at http://www2.hawaii.edu/~grace/emerg.html.
Pawley, Andrew. 1991. "How to talk cricket: On linguistic competence in a subject matter."
In Robert Blust (ed.) Currents in Pacific Linguistics: Papers on Austronesian Languages and Ethnolinguistics
in Honour of George Grace, 339-368. Canberra, Pacific Linguistics.
This segment posted 20 May 2008
Note added 12 June 2008
Minor changes made 23 June 2008
Conceptual Situations and Events
The other two components
Some might say that what I'm calling "the saying mode" of communication doesn't really belong in a sequence on the evolution of subject
matter--that it's not really so much a matter of what can be talked about as of how we talk about it. However, this mode of communication
is critical to language as we know it today, and omitting any discussion of it seems to leave the story very incomplete.
I suppose it might be argued that the saying mode really boils down to just a single innovation: having someone assume the role of
sayer--i.e., assume responsibility for the utterance. But that seems a very incomplete description. The huge difference between the
kinds of information that can be conveyed by
the full-fledged language that we know today and what could have been conveyed in the (hypothetical) evolutionary stage described in
the last
segment suggest revolution rather than evolution. However, I want to argue that all of the change could have been achieved through small steps.
Of course some--no doubt most--of the change has nothing to do with the saying mode in a narrow sense, but could hardly have progressed
very far without it. Anyway, I'll have more to say about these changes in what follows.
First, I need to say a bit more about how the saying mode works. According to my analysis (which comes essentially from my 1987
book The linguistic construction of reality) the saying mode has three components. One of these, conceptual situations
(and events),
was briefly introduced in the last segment. That segment was a discussion of how the subject matter of the beginning-to-evolve
language might gradually have expanded to include more reference to things external to the speaker--more on his/her "objective situation".
One might want to say that of the three components of the saying mode, the conceptual situation/event is the only one that's really about subject matter.
But that segment was concerned only with how such things might have gotten started at all, and of course they've had to undergo enormous elaboration to
permit language as we know it today. I won't attempt any speculations about the subsequent evolution of conceptual situation/events, but I will
try to describe more clearly what it achieved.
Some readers might find it helpful to think of the contrast between conceptual situations on the one hand and the "subjective" and "objective"
situations we were previously discussing as one between emic and etic. A conceptual situation is typically an objective situation--more
specifically an "effective objective situation" as described (actually or potentially) in a particular language. It's a particular depiction of
that situation; it's the particular situation as a speaker of a particular language has defined and structured it in that language--by means of the
resources of that language. A "conceptual event", of course, is a similarly depicted event.
Note that conceptual situations and conceptual events are entirely parallel; both refer to what our language enables us conveniently to
recognize, encode, and report as possible states of affairs or happenings. But I can't find any term that encompasses the two (a fact that's
been a source of frustration for me for years). Therefore, I'll resort to the abbreviation CSE to refer to a conceptual situation and/or event.
I've discussed CSEs in my book (Grace 1987). There are just four points about them that I need to bring out here.
1. The very existence of CSEs is dependent on a conception of reality--on what I call a "conceptual world". Being able to
conceive a CSE requires some assumptions about what kinds of elements--states, processes, acts, objects, qualities, and the
like occur in the world and about how these elements are interrelated.
For CSEs to play a role in communication, this conceptual world must be shared. I've talked elsewhere (e.g., Grace 1987, 2005) about the role of
evolving language in the construction of this shared reality. Such a reality conception is what has often been referred to as "knowledge of the
world", but "knowledge" implies an unjustified certainty for what is actually a system of assumptions (many [though not all] of which would qualify
as beliefs).
2. As CSEs became more elaborate and nuanced, it eventually became possible to model things that were internal to the speaker (subjective situations
and events). However, these represented subjective situations have nothing to do with the effective subjective situations (whatever
it is in the speaker's overall state that motivates her/him to execute a particular speech act). Anyway, the point is that we can no longer say that
CSEs only model the outside world; they now reach into the inside world as well.
3. The same objective facts can be represented in very different ways at the choice of the speaker. A single actual happening can be represented
as many quite different conceptual events. To illustrate with an example I used in Grace 1987, the sentences "A man sold the doctor a car" and
"An Irishman unloaded a broken-down 1981 Chevy on that sucker" might both be used to report the same event. (Or, alternatively, the buyer might
have been represented as the agent: "The doctor bought a car from a man").
4. Of course, such sentences aren't bare CSEs at all. They exemplify the saying mode with all its components intact; they're fully equipped
for saying something by themselves.(1) As I analyze it, communications in this mode involve three components: in addition to the CSE,
there are what I've referred to as "contextualization" and "specification of the condition of instantiation" or (more succinctly)
"modality". Although the latter two play essential roles in speaking as we know it, one might argue that only the CSE actually should be
counted as representing subject matter.
Still, it's probably true that these capabilities were all evolving at pretty much the same time. It's probably even true that advances in
each helped pave the way for advances in the others. However, they still appear to be logically distinct in such a way that the most fruitful
approach to discussing them is to analyze them separately.
I'll try to illustrate how the other two components work. Consider the famous (at least to some generations of linguists) example,
"the shooting of the hunters." Of course, what this particular expression was famous for was its ambiguity (were the hunters shooting
or being shot?) but the ambiguity would be a distraction here, so let's just stipulate that it's the hunters that are doing the shooting.
Anyway, the point is that "the shooting of the hunters" contains a complete CSE. However, to utter it is not to say something in the sense
intended here because it's lacking a modality component. Nevertheless, this expression is not a bare CSE; it still contains contextualization clues.
To get a bare CSE in this case we must remove the definite articles. Thus, the actual conceptual event is better expressed by something like "shooting
of hunters", or simply "hunters shooting".
What we're left with is a representation--a kind of depiction--of an event (or more exactly, kind of event) in which there are (plural)
actors who belong to the category "hunter" and who perform an act of the category "shoot". This, then, constitutes a general kind
of event--the "hunters shooting" kind of event--that the language makes it easy to conceptualize, and that is therefore readily available
to be used in saying something. This particular conceptual event is a very simple one, of course. CSEs can become much more complex; this
one itself could be expanded to, for example, "inexperienced hunters shooting at doe out of season with AK47s".
As mentioned above, this CSE lacks two components required for saying something. (Well, strictly speaking, I suppose contextualization is
not required for saying, but certainly the range of what can be said without it is pretty limited). Anyway, I'm going to discuss it first
for the simple reason that I find it hard to come up with very good examples without it.
Contextualization
I find it hard to imagine how any but the simplest CSEs might have been used in the absence of some means of adding contextualization and
modality information. Consider contextualization: A CSE without contextualization is like a picture without identification. Until we're
told that it's a picture of the murder scene, that the body was here, the weapon there, etc., it means nothing; it's only when we're given
that information that it acquires relevance. The same can be said of a map without labels--without our being able to tell what it's a map of.
Otherwise put, we need an answer to the (expectable) question: What are you talking about?
In the case of our hunters-shooting example, the deleted definite articles would presumably have provided part of the answer. The expression
with them restored (i.e., "the shooting of the hunters"), specifies that what we're talking about is some particular hunters (2)
whom the addressee is presumed to be able to identify (from either the context of the speech act--those hunters that we are or have been looking
at--or of prior discourse--those hunters of whom we've previously spoken). At least s/he is presumed to be able to identify them sufficiently
precisely for our present purposes--whatever these purposes are.
The other definite article--that preceding "shooting"--also has a contextualizing function. It might refer to some particular incident
or pattern of behavior in which the hunters in question were actors although it could simply assume a shared expectation that shooting
guns is a natural function of hunters--something to be taken for granted. In any case there would be the assumption that the addressee
would sufficiently recognize the allusion.
Contextualization clues depend on what is frequently referred to as "shared knowledge", and there are only limited ways in which this
"shared knowledge" could have been acquired. There is, of course, the kind of shared observation that exists when speaker and hearer
know they both were seeing--or hearing, feeling, smelling, etc.--the thing in question. Nowadays, for example, one person might say
to another "That guy looks like he's lost", speaking of someone visible to both.
However, full-fledged language permits us a much wider range of reference than that. In today's language, contextualization depends
particularly on prior discourse, by which I mean previous utterances (or writings). The most straightforward kind of case is where
speaker and hearer have discussed the thing in question before. Their previous discussions (oral or written) may be described as
constituting a discursive context for whatever is to be said. I expect you to be able to understand who or what I'm talking
about because we've talked about it before: I only need to provide a clue that will remind you.
But usable discursive contexts may be found much further afield: for example, I may have inferred from our prior discourse that
you have heard or read some particular thing. Nevertheless, a speaker very often just makes educated guesses about the discursive
context accessible to the addressee. What I know of your history just leads me to infer that you'll recognize a particular allusion.
Understandably, contextualization clues based on such inferences don't always work.
The whole matter of discursive contexts deserves a fuller treatment than I can give it. I did discuss my understanding of it in
a little bit more depth a long time ago (in Grace 1985). In any case, the relevant prior discourse can range from very
private to very public--from something uttered in private conversation to something one can assume everyone of similar ilk has
heard or read. We may think of the sum of such discourses that are in the ken of both speaker and hearer as constituting the
discursive context of the utterance. Every utterance (except, presumably, for cases of first contact between members of
different cultures) has some kind of discursive context, and this discursive context is available to provide contextualization
for the utterance--for the situation or event being depicted.
Anyway, the kind of language use that we know today depends very heavily on shared knowledge, and this shared knowledge derives
in very large part from a shared discursive background. The very existence of such a thing as shared discursive backgrounds seems
a radical innovation in the evolution of language: our concern here is whether it could have evolved naturally--by small steps--or
whether it required a lucky miracle. The answer, I think, is that it could have evolved gradually.
My guess would be that it came about in something like the following way. There were already continuing (or ongoing) relationships
between individuals who customarily interacted. Once individuals reached the point of addressing messages directly to specific
addressees about subjects unrelated to them, these messages would become part of the ongoing relationship just like other interactions.
As the potential subject matter expanded, the participants presumably came to remember previous communications just as they remembered
previous interactions of all sorts--from groomings to confrontations. For our purposes it's useful to analyze out the messages from the
rest of the ongoing relationship and think of them as constituting an ongoing discourse. Of course, the efficiency of communication
was greatly enhanced as it became possible to introduce a subject into one's message by simply alluding to mention of it in prior discourse
rather than having to provide it a full introduction. Thus, prior messages would be part of each individual's understanding of the relationship
just like other prior actions. They would be available to be used as a context for references.
Of course, even if something like this is the explanation for how ongoing discourses might have gotten started, it still doesn't explain
how one could go about identifying a referent within the prior discourse (e.g., filling the role played by the definite articles in our
shooting hunters example). I don't have any particular explanation to propose. I would imagine that once the amount of information that
had been communicated between two (or more) individuals began to be felt as significant, they'd have felt an increasing need for means
to make specific references to it. And, I'd imagine that some--presumably crude at first--means (verbal? gestural?) would have been found
to begin satisfying that need. Then (I'd imagine), these devices proved advantageous and they evolved rapidly in use and sophistication.
Although I won't undertake any suggestions about just what the steps might have been, I don't find it hard to imagine that the kind of
contextualization cluing that we know today evolved through gradual advances without any need for the intervention of miracles. It seems
safe to assume a gradual process until someone comes up with evidence that there was some advance that could only have been achieved
through a great leap forward.
Modality (Specification of a Condition of instantiation)
This is actually the key component of the saying mode. Nothing has been said until a speaker plays the role of sayer by
assuming responsibility for the specified condition of instantiation (of that event in that context). A modality is, in fact,
an essential element of any complete sentence.
It seems natural, at least to the language ideology of modern global culture, to assume that assertion is the prototypical modality.
A clear example of this modality is represented by the "assertion sign" in Whitehead and Russell's Principia Mathematica which in their words
is "required for distinguishing a complete proposition, which we assert, from any subordinate proposition contained in it but not
asserted." (Whitehead and Russell 1957: 8). Our expression the shooting of the hunters is what they describe as a "subordinate
proposition" designed to be contained in another proposition.
Thus by adding a modality we can, as Whitehead and Russell suggest, use our example expression to make an assertion--say,
"The hunters were shooting". In so doing we would be declaring that the "hunters shooting" CSE did indeed occur (was in
fact instantiated) in the indicated context.
But other modalities are also possible; most obviously a denial of the existence of the specified instantiation--e.g.,
"The hunters weren't shooting"--or an inquiry as to its existence like "Were the hunters shooting"? It seems possible
that the assertion modality only became defined in response to negative, interrogative, or both.
As we've said, modality is really the essential component of the "saying mode" of communication. It permits us to use CSEs in
messages to do things other than asserting that they exist, and, for that matter, even to make that assertion. Although I've
always used the term rather casually to include such things as contingencies, aspects, and maybe tenses, it seems to me that
the most basic modalities and most likely the first to be distinguished would have been assertions, yes-no questions, and negations.
Again it seems to me that the initial modality distinctions could have arisen by gradual steps. For whatever it's worth I'll offer a speculative story.
Suppose that in one of our ancestral communities there was an utterance (e.g., a vocalization) used to call attention to an approaching predator (as suggested by the vervets). Now suppose that individual a arrives, visibly upset, in the vicinity of b. The latter notes a's state, finds it disturbing, and wants to ask something like "What have you seen that's got you in such a panic"? The best s/he can think of to do is make a guess, and produce an utterance--a trial utterance representing a hypothesis. The hypothesis in this scenario is that a has seen an approaching predator, and one option for testing it--in fact the only one that occurs to me--would be to produce the utterance normally used to signal a predator threat and hope it will be interpreted as a question. (Even today this kind of expedient is used, but normally with some kind of question marker or intonation.)
Of course, suggestions about CSEs being uttered in the hope that they'd be understood as questions doesn't explain at all how different modalities came to be distinguished formally. The only thing I can propose as an explanation is the same as I proposed for contextualization clues. As communication came to include a wider ranger of subject matter and generally assumed a larger role in community life, there would have been an increasing need for the ability to specify modality--to make assertions or denials and to ask questions. And, I'd imagine that some--presumably crude at first--means (verbal? gestural?) would have been found to begin satisfying that need. Then (I'd imagine), these devices proved advantageous and they evolved rapidly in use and sophistication.
I might mention that one reason why I'm inclined to extend the concept of modality to such things as "would the hunters have been shooting if (...)"? is that I think of the modality distinctions from the very beginning as requiring the ability to conceive of what have been called different "possible worlds" or "mental spaces".(3) A CSE in fact represents a mental space, and our modalities simply specify the conditions of their instantiation.
NOTES
1. We sometimes speak as if the sentences themselves said something. That's not true, of course; the sentence is an instrument suitable for use
by a speaker for saying something.
2. What's suggested here is in fact an oversimplification of how language works. The expression "the hunters" doesn't have to refer to
individuals who meet the definition of the word "hunter". It could be used to refer to anything. For just one example, we might previously
have referred--sarcastically, humorously, or whatever--to some group of possibly very non-hunter-like individuals as "the hunters". (In such
a case the bare CSE would be something like "[unspecified actor] shooting").
3. This term "mental space" is due to Gilles Fauconnier. The earliest source that I'm aware of is Fauconnier 1985. I don't know the source of
"possible worlds".
REFERENCES
Fauconnier, Gilles. 1985. Mental spaces : aspects of meaning construction in natural language. Cambridge, MA: MIT Press.
Grace, George W. 1987. The linguistic construction of reality. London: Croom Helm
Grace, George W. 2005. A "Swiss-Army-Knife" Conception of Early Language Evolution. Reflections on the evolution of human language,
number 5. Internet WWW page at http://www2.hawaii.edu/~grace/knife.html.
Whitehead, Alfred North and Bertrand Russell. 1957 (1925). Principia Mathematica, 2nd edition. Cambridge: Cambridge University
Press.
This segment posted 20 August 2008
The main objective of this discussion of subject matter has been to provide some ideas about how the evolution
that led to what we know today as language might have been continuous and gradual. Of course, fully evolved
language is enormously different from anything within the capability of our mammalian ancestors. This, of course,
raises the question of what it was that made all of this change occur. What function drove the evolution that
brought it about? When I speak of a function as "driving" the evolution, I mean a function that was served by
behavioral innovations that were, consequently, selected for in subsequent generations.
A main concern has been to show that there is no need to assume this driving function to have been something
that emerged abruptly--in any sort of metamorphosis or "saltation" in the evolution process. I've particularly
tried to offer an alternative to any predisposition to assume that the driving function was conveying propositional
information. I've tried elsewhere--most recently in Grace 2007--to show the inconclusiveness of some of the arguments
for that assumption.
According to this alternative hypothesis the function that drove the evolution was enhanced rapport and empathy, and
that it was this function that drove the evolution from the beginning. In the previous segments I've described the
basic subject matter in terms of interpersonal relationships or "subjective situations". I think of the fundamental
message as being something like "Here's what I'm experiencing that I feel like sharing with you". But the larger point
is that the rapport between individuals and within the community was steadily enriched as the capacity for sharing
experiences and the scope and subtlety of the experiences that had become shareable grew.
But why would such enhanced rapport provide advantages sufficient to drive this evolutionary process? My hypothesis
is that the advantages proceeded from greater group cohesion with more frequent and more sophisticated cooperative activities.
One might think of these critical advantages as similar to those that led to the insect societies that have sometimes been
described as "superorganisms". The species in question include most famously bees, ants, and wasps (of the order Hymenoptera)
and termites (order Isoptera). The description superorganism refers to the analogy between an individual organism and an entire
colony of such insects in which the individual members of a colony have specialized functions reminiscent of the organs of an
organism. Humans have achieved some of the same advantages of such superindividual organization by non-organic means. The point
is that, in so doing, human society has become extremely different from any other primate society both in the number and the
interdependence of participating individuals, and that this difference as been crucial to our success as a species.
Of course, the differences between human and insect society are numerous and significant. We humans have a much more intricate
and constantly evolving system of roles, with genetic factors playing a quite limited part in selecting their occupants.
Nevertheless, the same metaphor of a higher-level organism where individuals play roles analogous to those of the organs
in a biological organism has some degree of applicability in both cases.
The first step in such evolution would have involved an expansion of the communicative repertoire--a larger repertoire
of messages becoming available. Interpersonal relations would still have constituted the central burden of the messages,
but with what could be conveyed about them becoming increasingly subtle and specific. Furthermore, the expansion would
have begun the first exploratory steps into information that was less narrowly personal. This enriched repertoire of
messages would have provided the basis for enhanced rapport between individuals. It would also have permitted an individual
to interact with a more diverse set of other individuals and probably led to networks with more participants and more complex
relationships. All of this should have resulted in enhanced cohesion for the entire social group, which would presumably have
led in turn to the coordination of activities of increasing complexity.
One of the eventual benefits of cooperative activities was specialization. Individuals could focus their attention on the
development of particular sets of skills and on producing, and ultimately designing, more complex tools.
Populations in which something like this happened would presumably have been able to exploit their environment in increasingly
advantageous ways and would have prospered and grown accordingly. Most importantly, the individuals who had proven most
skilled in these beneficial activities--effective communication being the most basic of them--would have enjoyed especial
advantages that would probably have included reproductive success. And thus the evolution would have continued, leading
among other things to the steadily-increasing range of subject matter that was discussed in previous segments.
Furthermore, the reach of organization has also been steadily increasing. Independent societies have, through conquest or
trade, been incorporated into larger cooperating networks to such an extent that today the entire world is in some measure
implicated in a single phenomenon known as "globalization". It's become more and more appropriate to think in terms of a
global society with its own rules and its own interacting parts, all with their own assigned roles.
In sum then, the hypothesis is that the traits that have made the species depend on elaborations of the underlying message
that I've roughly described as "Here's what I'm experiencing that I feel like sharing with you". One of the elaborations that
seemed particularly noteworthy was the increasing ability to bring the outside world (one's "objective situation") into
discussion. Another was the ability to conceive of different "mental spaces" and the emergence of the "saying mode". But
when matters are better understood, no doubt there will have been others that deserve equal or greater attention.
Anyway, it was these elaborations--and ultimately the capabilities that enable them--that led to our present highly advanced
organization on a superindividual level involving large units of highly specialized cooperating parts. And as a further
consequence we have acquired means that far surpass those of any other species to adapt to--and even modify--the affordances
of the physical environment.
REFERENCE
Grace, George W. 2007. Thoughts on reading Heryanto. Internet WWW page at http://www2.hawaii.edu/~grace/Heryanto.html.
This segment posted 28 January 2009
Introduction
The nature of humans' knowledge of language: Some prevalent assumptions
Evolution in knowledge of language
Evidence of lexical-type knowledge today
That the acquisition of such lexical-style knowledge seems straightforward enough
On the later KOL evolution
On the economy of KOLs
The question with which we began this sequence concerned the functions that drove the evolution of language--what it was enabling us (or, more to the point, our ancestors) to do. What advantages did the evolving language afford? In this sequence the focus was subject matter; the discussion has consisted mainly of speculations about how the range of subject matter with which our evolving ancestors were able to deal linguistically has grown. And that must surely be a major part of the answer.
It was Noam Chomsky who reminded us that a language in fact consists just of what he called the “linguistic competence” or the “knowledge of [the] language” of those who speak it. Certainly, what’s required to use language competently may indeed be conceived of as a kind of knowledge. In those terms, the uniquely human capacity for language that is our object consists of an ability to acquire this kind of knowledge plus, of course, whatever innate motivation impels us to do so.
What, then, is the nature of this knowledge? What kind of thing does a human have to know in order to be accepted as normally competent linguistically? The most objective measure of normal competence presumably would be the judgment of the speaker’s community, or at least the community in which s/he had acquired the language in the first place. The kind of knowledge required would then be defined as that possessed by the individuals who are generally accepted as normal speakers.
But what kind of knowledge would that be? I certainly can’t provide anything like a complete answer, but I think a good place to start might be by questioning the validity of two assumptions that seem to be taken for granted in much linguistic work.
First, there seems to be a kind of tacit assumption that monolingualism is the norm. Thus, in the prototypical case, knowledge of language (henceforth KOL) would be nothing more than knowledge of a particular language. Once the conception of KOL as knowledge of particular languages is established, it’s easy to extend it to multilinguals simply by assuming each multilingual KOL to be nothing more than a collection of separate single-language KOLs.
This assumption seems clearly misguided. At least, it is if what we’re concerned with is really the knowledge required to use language competently, and if we accept the community’s judgment about competent use. It has frequently been pointed out that it is monolingualism that is exceptional in the human experience, whereas multilingualism is the norm. But multilingual experiences can be infinitely varied, and multilingual KOLs surely could not all be expected to conform to any single model.(1)
A second assumption has to do with the nature of these postulated single-language KOLs. The standard assumption seems to be that what we know when we “know” a language consists mainly of a kind of grammar. Moreover in this grammar, it is often syntactic relationships that are assumed to have the greatest significance. In fact, this has led rather naturally to a further assumption that the uniquely human capacity for language of which we spoke boils down to a capability to entertain certain kinds of syntactic relationships.
I believe this assumption is also misguided. In fact, as I’ll explain later, I think the structure of KOLs is more lexicon-like than grammar-like.
One problem that I have with the grammar hypothesis is that it seems to require an evolutionary leap. It makes our KOL appear fundamentally different in kind from anything within the capacity of any other species including those most closely related to us. It seems to follow from this view that there must have been at least one abrupt change of direction in the course of our evolution, and that such a change must have required a mutation or at least some kind of transforming innovation to bring it about. And of course, a main point of this entire sequence on the evolution of subject matter has been that no such abrupt transformation needs to be postulated.
Of course, the kinds of knowledge-of-language that individuals have been capable of have certainly changed during the course of the evolution. In fact, such changes would inevitably have been a critical aspect of this evolution since the inclusion of any subject matter depends on speakers with the competence to make use of it.
Our species presumably set out on its independent evolutionary path with the kind of capacity for representing subject matter that we retained from our mammalian ancestry. The subject matter at this stage seems to have consisted essentially of expressions of feelings, especially those involved in interactions with others. We’ve assumed that even at this point the repertoire can be interpreted as a fixed set of distinguishable signals, each of which signals a distinct kind of feeling.
At a later stage when we were dealing with the increasing range of expressible subject matter, the concept “subjective situations” was invoked to replace the vaguer “feelings”. Of course, subjective situations are still ultimately feelings, but feelings that can be influenced by multiple different factors at once.
In the same segment I spoke of “some kind of knowledge of associations between subjective situations and linguistic expressions”.(2) The notion of subjective situations matched with linguistic expressions refers to the dictionary-like model that I’ve proposed in various forms in several other places--Grace 1987, 1989a,b, 1993, 1997. One’s knowledge of language, according to this model, consists of the inventory of linguistic expressions (or their pre-linguistic antecedents) each of which is paired with (and in effect, defined by) the situation or situations to which it is an appropriate response.
But this description doesn’t give a complete picture of what I imagine to have begun developing even from a quite early stage. What I envisage is a kind of feeling (“Sprachgefühl”) for what will be understood of our intentions if we perform a particular speech act, and for what is to be inferred about someone else’s intentions from his/her own speech acts.
Still we might, as a first approximation, think in terms of a sort of lexicon in which linguistic expressions are paired with subjective situations. Of course such a simple model is obviously too limited to account for the creativity of later stages, but I believe that KOLs still remain much more like a lexicon than like a grammar--that this structure characterizes not only the KOLs of the earliest stages of the evolutionary process, but also the framework on which language as we know it today is built. In fact, what led me to this model in the first place was evidence that we still regularly resort to direct knowledge of such dictionary-like matches today.
A number of things over the years contributed to convincing me that lexicons were better models for our KOLs than grammars. It’s apparent in my 1981 book that I’d pretty much reached that conclusion already. For example I said, “…it is obvious that the characteristics of man’s linguistic competence are fundamentally different from the characteristics of the usual linguistic descriptions. The economy of linguistic description is radically different from the economy of the mind.” (Grace 1981: 17). And see also the chapter entitled “What is there to be calqued?” (ibid. 43-54).
What I was certainly already convinced of at that time is that much more of humans’ knowledge of language is holistic--stored in memory--and much less analytic--a matter of general rules--than our theories had presupposed. I arrived at the concept of subjective situation in my efforts to make sense of the fact (noted in the “Situations” segment) that when presented with a speech act in a source language, bilinguals regularly are able to judge whether or not a proposed linguistic expression in the target language is what speakers of that language would say “in the same situation”. (For further discussion, see Grace 1987, 1989a).
However before that, there had been clear evidence that we associate linguistic expressions with situations (in some sense of the word). One thing that struck me as a pretty startling discovery came from the “grammaticality judgments” we were all trying to make in the early days of generative grammar. What I found myself doing when asked if a certain questionable string was “grammatical” for me was try to imagine a situation where the string in question might be appropriately uttered.
Everyone else with whom I ever discussed the matter reported the same experience. I found this significant as well as startling because these judgments had nothing to do with the grammatical structure of the linguistic expressions--it was their match with a situation that mattered. (I must admit that nobody else seemed to think this mattered--I never understood why).
It isn’t hard to see how knowledge of such connections could be acquired even today (regardless of whatever other forms of knowledge were being acquired at the same time). All that would be required presumably would be enough storage capacity along with a natural impulsion to interact with others in ways that elicit favorable responses. The responses of others to the learner’s behaviors could be expected to reinforce some behaviors as opposed to others, and it would quite soon become apparent that different behaviors are suitable for different situations.
At some stage the learners would become aware of their capacity for imitating--to varying degrees of approximation--the behavior of others. They then could recognize these behaviors as potential models and probably pay increasing attention to them, their contexts, and the responses they elicit.
Eventually a stage should be reached where it would become apparent that the “situations” to which the behaviors were matched were not objectively observable in their entirety--that they depended in part on the speaker’s internal state--his/her feelings, motivations. In other words, the learner would come to know, consciously or unconsciously, that the relevant “situations” are what we’ve been calling subjective situations.
The result of this evolving knowledge would thus be a kind of repertoire of situations--each associated with its own socially appropriate behavior(s). Although not all of these behaviors would necessarily involve speaking, many certainly would.(3)
Of course, as time went on our ancestors became capable of greater perception and discrimination of elements of the external world and increased introspection into human emotions and motivations. As this happened, the knowledge required for acceptable language use became increasingly complex. The expansion of our conceptual worlds both necessitated and depended upon extensive elaboration of the associated KOLs.
But it’s hard to be at all specific about the KOLs of more fully-fledged language. One major problem is that there isn’t a single KOL for any language(4), much less for any multilingual community, because each speaker constructs and maintains his/her own KOL, and no two are identical. Furthermore, these individual KOLs are themselves never entirely fixed; they are always works in progress.
All I can really attempt, then, is to talk about some of the factors and considerations that would probably have played a role in opening and elaborating the dictionary-like structure of the early KOLs.
For one thing, the lexical entries would have had to be rather loosely defined even from a very early stage. Any of the reference situations is conceived in general terms so that it might be thought of as actually specifying a set of situations that are equivalent--“equivalent” at least in the sense that they are all represented by the same linguistic expression. In fact, such a blanket conception is obviously necessary in the case of the “situation” entities since no two actually-occurring situations can be identical in all details. (Of course, this means that when dealing with a given actually-occurring situation, the language user has the responsibility of judging its equivalence to any particular lexical entry).
But even here I need to qualify. Language users constantly create new equivalences. Or it would be more accurate to say that they perceive likenesses that may not have been perceived before--and these likenesses, if accepted, become equivalences. (This capacity for perceiving likenesses, of course, is the basis of metaphor).
The digital nature of linguistic structures, on the other hand, does make it possible for a linguistic-expression entry to be so precisely specified that any actually-occurring expression either is or is not identical to a particular lexical entry.(5) In practical application, however, this entry (what it is appropriate to say in a specified situation) is also most often conceived in a general way with quite a range of non-identical expressions effectively being interpreted as equivalent.
And to complicate the picture further, the match between situations and speech acts is, of course, not precisely one-one anyway. Nor is it precisely many-one or many-many. It’s a more complex kind of mapping. If we start from a particular speech act (i.e., assume we’ve observed the act, and are trying to infer the subjective situation), there can often be multiple situations that for our immediate purposes would be equivalent.
And the perspective from a particular subjective situation is likely to be parallel: there are likely to be more than one linguistic expression that are approximately equivalent for our purposes.
This suggests a kind of problem that the language user must deal with. It also suggests a mechanism by which creativity was (and is) exercised and by which complexities were added to the basic lexicon. Imagine, for example, that we find ourselves in a subjective situation that leads us to want to speak (that is, to give expression to our subjective situation). However, this situation isn’t exactly like any we’ve experienced before--the feeling we want to express is different.
Or to put it in terms of our dictionary model, we find two lexical entries--i.e., two reference situations--each of which partially matches our actual situation, but neither of which matches it exactly. To utter the linguistic expression associated with either reference situation would be at once suggestive and misleading.
We realize that the situation might be most accurately represented by some combination of the two reference linguistic expressions. More specifically, the best fit would be achieved by analyzing and recombining the two in such a way as to suggest the way in which our ostensible subjective situation corresponds to each of the reference situations--in other words by designing a kind of compromise linguistic expression in such a way as to represent the relevant parts of each reference situation. The capacity to do this would, of course, require an ability to analyze the linguistic expressions as well as the situations. This is a critical step, of course, but too big a subject to go into here.(6)
What I’ve been able to figure out so far about how it might have been achieved can be found in my discussion of “the emergence of analytic processing” in Grace 2008. The hypothesis presented there is that some expressions could be perceived as divisible into different parts, and that when that occurred each part sometimes came to be associated with a different part of the overall meaning. This, of course, would have led to the potential for recombining parts to create novel expressions.
Of course the kind of sloppy dictionary model that I’m proposing as the way our knowledge of language is structured is about as inelegant as any model could be. It contrasts pretty directly with a primary goal that has been pursued in linguistic description. This goal has been conciseness; some critics have described it as conciseness as measured by the amount of paper required to present it. It aims at a maximum reliance on the application of general rules and a minimal reliance on memory.
My excuse is that the brain doesn’t work that way. It isn’t elegant--at least not in that understanding of elegance. As I tried to point out in Grace 1981, the economy of the human brain--of cognition--seems very much at odds with the economy of linguistic description.
NOTES
1. The most favorable scenario for creating a KOL composed of nothing more than separate single-language KOLs would presumably be one in which the individual has acquired and used each language in an entirely separate setting. For example, one might imagine a person living what are effectively two separate lives--some periods in community A speaking language A and others in community B speaking language B--and experiencing little or no need ever to integrate the separate experiences. By contrast imagine one living in a fully bilingual community in which code-switching was freely practiced and it was common not to remember what language had been used. What these two persons would need to know in order to be perceived as normal language users would surely be quite different.
2. I should probably point out that this implies an assumption that I’ve made pretty much throughout all of my discussions of the evolution of language: that from very early in the process it would have been possible to recognize some kind of discrete, repeatable signals that were known and used by the community.
3. In this conception the speech acts are not accorded any special status or segregated out from other behaviors. The reason is that I’m uncertain how instinctive (or how universal) it is to conceive of what we abstract out as “language” as constituting a separate compartment of knowledge, or use of language as a separate compartment of behavior.
4. It is true, of course, that in many societies there is an assumption that there is a canonical version of what is effectively the KOL for their particular language, and that what the individual speakers do is attempt to duplicate it. But this is at best a (perhaps) useful myth.
5. Although this is true of linguistic expressions, it isn’t true of the speech acts through which they are deployed.
6. Does the capacity for analytic processing mean that these individual KOLs contain grammatical rules? My tentative answer would be that individuals differ pretty widely and that the environment in which the KOL construction took place has a substantial effect. For example, in the environment in which I live caregivers are likely to think in terms of preparing the child for literacy, and to have ideas about what is to be acquired that include such concepts as words, letters, and grammatical correctness. Some of their ideology is likely to be transferred to the acquirer and to influence the final product. But beyond that, I don’t want to speculate.
REFERENCES
Grace, George W. 1981. An essay on language. Columbia SC: Hornbeam Press.
Grace, George W. 1987. "What they would say in the same situation". Ethnolinguistic Notes, Series 3, Number 31. Also (1996) Internet WWW page at http://www2.hawaii.edu/~grace/eln31.html
Grace, George W. 1989a. The association of situations with linguistic expressions. Ethnolinguistic Notes, Series 3, Number 35. Also (1997) Internet WWW page at http://www2.hawaii.edu/~grace/eln35.html
Grace, George W. 1989b. Recognition strategy and analysis strategy in language use. Ethnolinguistic Notes, Series 3, no. 37. Also (1996) Internet WWW page at http://www2.hawaii.edu/~grace/eln37.html
Grace, George W. 1993. What are languages? Ethnolinguistic Notes, Series 3, Number 45. Also (1996) Internet WWW page at http://www2.hawaii.edu/~grace/eln45.html
Grace, George W. 1997. Linguistic Change: 5. The individual's knowledge of language. Ethnolinguistic Notes, Series 4, number 6. Internet WWW page at http://www2.hawaii.edu/~grace/elniv6.html.
Grace, George W. 2008. Reflections on the evolution of human language: 2. The emergence of analytic processing. Internet WWW page at http://www2.hawaii.edu/~grace/emerg.
This segment posted on 27 June 2009
To go to other places in this website, click on one of the cells below
Put on the Web 18 March 2008
Minor changes made 23 June 2008
Latest addition 27 June 2009
Links to notes added 28 June 2009
© Copyright 2008 George Grace, Honolulu Hawaii. All Rights Reserved
