Causal Reference as the Intersection of Idiolects

Maxim Lebedev

 

Ρausal theories of reference were introduced as a way of understanding how proper names refer. According to the accepted views, reference can be done by three ways.

1.   In those (rare) cases, when our access to the examined thing is immediate, we can refer to things directly, fix the label directly on the thing. This use of names corresponds to the Russelian conception of logical proper names. Such privileged access to things is rather an exception, and as soon as we take some distance from our referent in space or time, our access will presumably become indirect.

The 'indirect' reference can be achieved by one of two ways:

2.   We can 'fix the reference' of ordinary proper names and names of natural kinds using definite descriptions, where the referents of the respective terms are the individuals and kinds denoted by the definite descriptions.

3.   Our use of proper names or common nouns can have a certain causal origin, where the referents of the respective terms are the individuals or 'kinds' that figure in some important way in these causal origins.

On last two ways of referring we use the respective terms to refer without knowing to what we are referring. We don't have to know the nature of these referents, while this approach to language is still purely extensional.

Causal theorists argue that the theory of descriptions cannot give a satisfactory account of how names refer to things. In their opinion, they offer a more adequate explanation of how names pick up their referents. In the classical theory of descriptions the name N designates object x, being used by an agent S if x uniquely satisfies all or majority of such predicates F that S would agree with 'N is F'; the name determines the designation through the criterion of adequacy. But in many cases the information which we attribute to specific individuals may be wrong. That is why the theory of descriptions could be supplemented by including a social element. According to such alternative, names should be considered as playing a referential role within the limits of linguistic community, and the descriptions, which are linked to a name, should be determined by the standard beliefs concerning the named individual. It must not necessarily be achieved by averaging the beliefs of linguistic community concerning this individual: the role of some descriptions can be increased — for example, of those belonging to such members of linguistic community who have the best knowledge of the individual in question ('experts'). Then, the choice cannot depend entirely on properties of the individual; the causal element should be included in any successful theory of reference.

According to the causal theory, if I am not the first one to use a specific name, then the referent of my use of this name depends on the use of this name by the person from whom I (causally) have learned it. And if this person herself was not the first user of this name, then the referent of her use of this name depends on the use of this name by another person, from whom she has gained the knowledge of this name, and so on to the initial 'baptism'. The introduction of a name can be considered simply as the act of labeling or assignment to a name of a certain meaning of the definite description (for example, the man who is standing before me now), or as the act of 'fixing the reference' of a name through the definite description. The referent of a name in each given case of its use is the function of its causal origin.

Causal theories were reasonably criticized as too uncertain in the specification of relevant causal chains which determine reference. In order to make our presentation more specific, we have to bring counterarguments, involving alternative causal chains (see Fumerton 1985: 121-29). So what are the ontological grounds for these causal chains? What, from such point of view, does specify the fact that the individual can be causally identified in intensional contexts — is it causal connection

If we use the notion of 'convention' in articulating the conditions of reference, it entails the mechanism of translating the individuating elements of convention through the set of individuals, in time and space. But how can such mechanism work? The answer here would be requiring additional consideration; if such mechanism is understood as causal, then the question remains, whether we can apply such notion of causality to the theory of meaning in general and the theory of reference in particular without claiming that the causal history of the term must be unique. What kind of causality could it be, if not of a Kripkean kind?

First point that we have to make here is to tell what actually are the relata of the considered relation? And, accordingly, what is the ground of causality? Do we use a term with this reference and not with another one because of the way the object is — or do we do it because this term used to be used with this reference earlier?

We can think of both answers. Then the causality in referential causal chains can be discovered in two ways:

1.   paradigmatically, as causal connection between the name and its referent;

2.   syntagmatically, as causal connection between the previous use of the name and its next use, i.e. between two events of usage.

Generally, causal theorists apply the syntagmatic model. The paradigmatic one can to some extent be present as an underlying account, but it hardly can be postulated explicitly: its exclusive acceptance as it is would lead us to some kind of naturalistic concept of meaning. It would involve the criterion of strict determinacy of linguistic meanings by facts or objects of reality. The paradigmatic causal relation here would embody the realist claim of a separate fact as truth-maker of each true proposition.

Syntagmatic and paradigmatic models are interconnected in the same way as are interconnected these two aspects of linguistic analysis reflecting the distinction between two forms of existence of linguistic elements — in the system of language and in the text (speech). Within the syntagmatic approach the elements of language are not considered as interchangeable but as ordered in position. In the paradigmatic approach the relation between any two elements is the relation of opposition of linguistic elements in their associative connections; in the syntagmatic approach it is the relation of coordination or comparison of linguistic elements in their linear connections. Differentiation of elements has absolute character in the paradigmatic model and relative — in the syntagmatic one: in some positions elements may differ but in some others they can coincide partially or completely. And unlike positional ordering of formal elements, semantic elements are ordered contextually ('position' and 'context' being key concepts in linguistic syntagmatic accounts).

 

 

 

Paradigmatic model

Syntagmatic model
Basic relation

Opposition

Coordination

Ordering

Vertical (interchangeability)

Linear
(positional for formal elements and contextual for semantic ones)

Differentiation

Absolute

Relative

 

With this regard, causal theories did not explain so far, how names can refer in language: what they do is explaining, how names can refer in speech — and to stipulate that this is the same would be too strong an assumption even from the point of view of the conception 'meaning as use'. To provide meaningful answers to questions of the paradigmatic theory, the use of the concept of syntagmatic causality in the theory of reference should give the theory of ordering of names, and, more specifically, not only mutual ordering of the individual uses of the same name, but also of the contextual ordering. With this regard the approach of Putnam (see Putnam 1975) is more promising than the approach of Kripke, as it does not limit the determination by the event of the initial use, and assumes further rε-categorizations. However in Putnam's theory, too, causality remains rather a general metaphor (probably fundamental).

It seems that the involvement of a more definite concept of causality could help to make clear how a causal theory could provide an account not only of such and such uses of linguistic expressions, but of the fact itself that names can refer. Thus, to clarify the concept of causality we could do better using more elementary concepts, which would allow us to handle problems of reference within the relatively clean ontology, without counterfactuals. That leads us to a metaphilosophical question — how fundamental is the concept of ‘cause’ in reasoning, and how do we come by it?

We can apply here the account of causal explanation and causal relevance developed by Wesley Salmon, taking as primitives the notion of a process and that of a spatiotemporal intersection of processes (Salmon 1998). The aim of this account is to distinguish

The basic idea is that an intersection of two processes is a causal interaction if both processes are modified in the intersection in ways that persist beyond the point of intersection, even in the absence of further intersections. A process is causal if it is capable of transmitting a mark — that is, if it is capable of entering into a causal interaction.

Let's consider the usage of the name as a process, and moreover, a causal one, i.e. such that can enter into causal interactions. With regard to the account of reference we can consider the causal interaction either paradigmatically or syntagmatically.

1.      Paradigmatic causal interaction is such a meeting of a term with an object in a situation of the usage of the term, after which its reference changes. These are cases of naming something with already existing expressions, or when we realize that we have no more referents for some of our terms, or we discover new essential features of something (which don't change our idea of the extension of the term) etc.

2.      Syntagmatic causal interaction is such 'comparison' of or 'adjusting' the situations of the use of a term — the previous event of usage with the next one — after which there is no change of its reference. We use the name with the certain reference because we are aware of the situation of its previous use with this reference.

Both paradigmatic and syntagmatic variants can be considered both synchronically and diachronically. The process of causal interaction can be considered as developing in time (or, say, for formal languages, in some conceptual space) or as consisting in the verification within the limits of a synchronous cut of language. However, we are still supposing that the causal chain of comparisons of situations of the term's use would lead us back to the situation of 'meeting'. In natural languages we are practically never able to follow this chain up to the beginning. There are some exceptions like specific terms introduced by scientists and names invented by fiction writers; but they do not present a counterargument here. If we could cleanse our language of theory-laden terms, starting with the recently introduced ones, like names of elementary particles, then continuing through 'mass' and 'impulse' to 'element' and so on into the prehistory of language formation, we would end up with nothing meaningful (Van Fraassen 1982: 14). Moreover, extending this idea of van Fraassen on ordinary language, we shall see, that we use such terms as 'quixotic' or 'snob' regardless to whether we are or are not aware that we do it courtesy to, accordingly, Cervantes and Thackeray. It is due to the fact that both those terms, for which we cannot follow the causal chain back to the beginning, and those for which we are able to do it, will be governed by the same linguistic rules, both synchronic and diachronic. That is why, in particular, we can successfully use causal theories to explain historical and geographical features of functioning of natural languages. But it is important to stress that causal theorists will suppose it possible to follow the causal chain from each individual event of the use of the term up to the event of 'meeting' in a certain referential horizon. The syntagmatic account of causal reference is based on this possibility.

Thus, we have two necessary preconditions for a successful reference:

(a) presence of a substantial change of referential meaning (including its initial occurrence — 'first baptism'), and

(b) absence (and even impossibility) of substantial changes in referential meaning through the sequence of individual events of using the name (usage 'salva referentiae').

Therefore, meanings, including new ones, can be considered as marks of causal interactions, showing the changes kept through the usage of the name (during the use of language) by an individual. It is important to stress that the causal processes here are supposed to be idiolects — the individual ways of usage of the natural language as a system.

The possible consequences bring us to distinction of paradigmatic and syntagmatic causality and, more widely, to the way, in which meanings function in natural language. Consider an example. Suppose that the American University in Venice has a research position, and the President of AUV tells the human resource officer to hire for this position a Cantabrigian. The HRO did not know the word 'Cantabrigian' before, but here it means a graduate of Cambridge. The HRO acquires this term and since then starts to use the term 'Cantabrigian' to denote a graduate of Cambridge. She can use then this term while talking to other people who can acquire it in turn, etc.

Her idiolect, that is her individual way of using the system of English, can be considered as causal process. It is causal because it is capable of receiving and transmitting marks, i.e. of entering into a causal interaction. The interaction of the HRO's idiolect (I2) with the idiolect of the President (I1) was held in the point of fixing the reference of the term 'Cantabrigian'. This intersection of two processes is causal interaction because both processes have received steady changes at their crossing. The use of the term 'Cantabrigian' acts as a marker of these changes: the HRO has acquired this term and she will use it on with the same reference, which it has in I1.

However, for the crossing of idiolects could be considered as causal interaction, both processes should receive steady changes. If in I1 no change happens, it could seem that we cannot consider this interaction and processes as causal. To see why we can (and even should) do it only in that case, we need the distinction between paradigmatic and syntagmatic causality in reference. From the paradigmatic point of view, here we should consider as causal such an interaction which would result in the HRO beginning to call a graduate of Cambridge 'Cantabrigian', and the President starting either to name graduates of Cambridge differently (say, Cambodians) or using the term 'Cantabrigian' with another reference. Obviously, such model would be absurd, as neither learning nor using language would be possible therefore. But while in I2 they had the situation of 'meeting', in I1 they had the situation of 'comparison' or 'adjusting'. The syntagmatic causal relevance for successful reference in I1 will consist in the absence of substantial changes while transmitting the referential meaning through the sequence of individual events of the use of the term. If the President will call a graduate of Cambridge 'Cantabrigian' even after the HRO started using this term with this reference, too, then we have to admit that the interaction of our individual usages of language is a causal one, that these processes of usage are causal, and that references marking the term 'Cantabrigian' in I1 and I2 are successful.

Success of causal reference at the intersection of idiolects

 

I1

I2

Type of causality

Syntagmatic

Paradygmatic

Reference

Remained the same

R1(I1)=R2(I1)

Changed

R1(I2)¹R2(I2)

where R is the referent, i.e. the relatum of the relation of reference: R1 — before the causal intersection, R2 — after it.

Then we can say that when idiolects I1 and I2 intersect, the transfer of certain referential meaning from I1 to I2 (e.g., in the situation of learning the language) can be considered as the causal mark left at the point of their intersection. The syntagmatic causal relevance for the successful reference in I1 will then consist in the absence of substantial changes of this referential meaning through (all) individual events of the use of this term in I1. If the term keeps its reference in I1 and receives one in I2, we have to admit that the interaction of two individual processes of the usage of language is a causal interaction, these processes are causal, and the reference of this term in I1 and I2, which marks the intersection, is successful.

Consider a counterexample. Suppose we had a different plot. When the President has informed the HRO that she should hire a Cantabrigian, the HRO decided that the University needs a specialist to count bridges in Venice, and found such a professional. Since it so happened that this researcher got her qualification from Cambridge, the position was filled nomologically correctly, and the HRO got no disapproval. Then she would start to use the term 'Cantabrigian' referring to bridgecounters and not to Cambridge graduates. Since it is not a successful reference in I2, it becomes obvious that the syntagmatic criterion should be extended.

Then for the interaction of idiolects could be considered as causal, the requirement of the use salva referentiae not only in I1, but also in the conjunction I1&I2, is necessary:

($R)(R1(I2)¹R2(I2)&R2(I1&I2)=R1(I1)=R2(I1))

The idiolect, as any linguistic phenomenon, can be considered doubly: synchronically and diachronically, i.e.

·        as a set of all terms and of all rules governing them used by an individual, and

This distinction is linked to a more broad one — between such presentation of language, according to which individual speech acts are but occasionally instances of the steady norm, and another one, according to which language is a creative process of generating texts. Accordingly, existing approaches to the analysis of meaning can be divided depending on their account of language as a sign system. The same phenomenon in language can be considered from two points of view:

·         statically, when we ascertain the presence of the linguistic phenomenon in question and of its distinctive attributes;

·         dynamically, when we are keen to identify or reconstruct process as a result of which this linguistic element is generated, or to find out, transformation of which elements it is.

In the former case we consider the analysed phenomenon as an immediate givenness, in the latter — as an inference from certain elements taken for initial and as a consequence of certain operations with them.

Thus, to build the concept of causal relevance that we need, we must choose for the role of the causal process the notion of idiolect treated not only diachronically, but dynamically, according to the approach to the language not as to the static system of signs, and as to the dynamic process of their use.*

So, the causal process in dynamic causal model of reference can be:

1.   the history of the use of the certain term in the certain idiolect (diachronic approach);

2.   the process, which result is the present use of the certain term in the certain idiolect (synchronic approach),

where causal process in the narrow sense of causal reference is rather the latter one.

We can see then that in the given context it is more preferable to consider references in an idiolect, than in natural language as a whole (approach peculiar to the traditional causal theories of reference) even in case we treat language not as a system of signs but as the process of functioning of this system. In the latter case the referential causal process could be considered as

1.   the history of the use of the certain term in the certain language — for example, English;

2.   the process, which result is the present use of the certain term in the certain language.

So, with the causal process of this kind, if we have the sufficient foundation to consider the appropriate individual belief of a kind 'R(I1)=R(I2)' as supported by explanation and if we have the well supported thesis about existence of the convention on reference in the relation R(I1), then we are warranted for the claim of causality for the marker, which now represents an individual belief (supported by causal explanation) that R(I1)=R(I2). If, then, we accept the condition of causal process for both participants of interaction, the situation of the use of language as a whole also can be described as a process of interaction of idiolects and as causal process, since each 'meeting' with a new idiolect, if it is marked in appropriate way in this idiolect, is marked as well in the common referential convention by occurrence of a new element — of the new individual, sharing this convention. The reference to the world will be implemented thus by means of support by other references of the same system, of which the causal process consists.

The use of the criterion of support by the system of references is aimed at grasping the intuition, most relevant for theories of meaning, according to which the situation of valid use of expressions of natural language has the form of a situation of existence of the agreement about their use. For the referent of the name could be considered as defined, the members of linguistic community should share a certain convention, within the framework of which the certain ways of individuation connect the name with the objects in the world and with experience of the individuals. The theory of causal process reference provides an account of how the linguistic convention is established and supported: the referential meanings are transferred through the interaction of two individual languages. Thus, the reference to the world is implemented by being supported by other references of the same system. Fixing and supporting references is the result of the interaction of idiolects, and the linguistic convention, which is in effect valid in certain linguistic community, is composed of events of such interaction.

 

Institute of Philosophy of the Russian Academy of Sciences

Moscow, 129110, Russia

lebedev@philosophy.ru

References

Fumerton, R. 1985. Metaphysical and Epistemological Problems of Perception. Lincoln and London: University of Nebraska Press.

Putnam, H. 1975. The Meaning of Meaning. In his Mind, Language and Reality, vol. 2, 215-71. Cambridge Mass.: Cambridge University Press.

Salmon, W. 1998. Causality and Explanation. New York – Oxford: Oxford University Press.

Van Fraassen, B. 1982. The Scientific Image. Oxford: Oxford University Press.

 


* Thus, strictly speaking, there are no obstacles for synchronous consideration. According to which of the dynamic aspects of language we want to describe — historical or functional, we can talk about dynamic models of a synchronous condition of language, on the one hand, and dynamic models of its diachronic state, on the other hand. In the synchronous dynamic model of language the generation of speech is perceived by the speaker as such an instantaneous process, all components of which exist for her simultaneously. The speaker has at her disposal a certain set of linguistic means, and the dynamism of language is shown mainly in the choice of linguistic means from the available set and in their organization on certain rules with the purpose of construction of meaningful texts. Dynamics of language are also manifested in the possibility to construct the texts of any necessary complexity of the rather limited set of linguistic means and in the deducibility of some elements from other ones.