LC3cA35

The Genesis of Symbolic Thought

Learning, in General


Let us begin by going beyond a “stimulus-response” couple to a stimulus-response arc. That “arc,” represented as a link between input and output (more generally stimulus — which may be entirely interior — and response — which may be entirely interior) is the site for attachment of interventions. In a simple case (e.g. Meltzoff’s experiment), the output of the SR arc mimics the input.

Interventions are, at first, interruptions of process, because of some sort of disruption (types might be failure, confusion, discordance, etc.); in this case, the links become sites for attachment of problem descriptions. When discriminations occur, these interrupted-links become the loci of extensions to the SR arc; how do the discriminations occur ? They follow Sussman’s formulation: problem descriptions are converted into prescriptions for change by local structure modification agents. (Minsky’s B-brains are intended to be capable of this functionality.) After a discrimination, the interruption has been repaired and is now an intervention in the preceding structure. Every link can be interrupted and the development of interventions occurs everywhere. As intervention-extended networks grow out of SR-arcs, they become slower in processing, more confusing for B-brains to manage, and ultimately, too complex for B-brains to change (this means they cease being capable of learning).

As these networks (societies of agents) grow, they compete with each other. Note well, the simplest processes (SR-arcs) still compete with them, and this can lead to the later replacement of a well established complex society of agents by a later developing but better fitting simple society.

Language-Specific Learning Theory


Following Peirce, we schematically represent three kinds of ways in which signs are involved with things signified: Iconic signs recall things signified, indexical signs indicate or “point to” things signified, and symbolic signs name things signified. (“Names” here implies conventional assignment of reference, variable by society and language groups.) The main issue to be explored is relations of the three kinds of signs, among themselves, and the way their interactions can be seen to explain important linguistic and psychological phenomena.

Iconic Signs as Fundamental Beginnings
In the simplest case, iconic signs are pristine SR arcs; they remind individuals of the things signified in that their intepretant recognizes No Significant Difference (NSD) between the icon and the thing signified. The interpretant for such an iconic sign is no more than a K-line which responds to the stimulus. The internal representations of the external sign or e-sign (which has in this case been taken as an iconic sign) becomes associated with the K-line; this association creates a change to the K-line where the counterpart “e-sign related modification to the K-line” serves as a personal-sign, or p-sign. This associated p-sign is used expressively to produce a vocalization intended to indicate the thing signified, e.g.:

  • the e-sign /”Scurry”/ used by family to refer to the dog [Scurry] is interpreted by Peggy to be associated with the dog [Scurry].
  • Peggy’s K-lines involving [Scurry] become associated with her p-sign for [Scurry] which, in her vocal expression, is manifest as /cul/di/.
  • when Peggy uses her p-sign { /cul/di/} to signify [Scurry], she is using the term as an indexical sign. This expressed p-sign functions as an indexical sign if and only if others recognize what it indicates.

How P-signs Become Indexical Signs
Knowing that Peggy expresses her p-sign for [Scurry] as /cul/di/, some people use the sounds /cul/di/ to refer to [Scurry] in the attempt to communicate with Peggy. It works. In this special case, the infant’s p-sign functions as the personal part of an indexical shared sign (s-sign) because the local Society accommodates to the infant. Consequently, association of the p-sign with the [Scurry] k-line is strengthened. The use by another person of /cul/di/ is not a p-sign itself (for them), but a transient, symbolic e-sign referring to the same entity [Scurry].

When the infant modifies the expression of her p-sign to accommodate to the Society, the p-sign becomes an indexical sign through a different process, as follows. Other people use the e-sign, e.g. /”Scurry”/ to refer to [Scurry]. Interpretants need to be developed to relate Peggy’s perception of /”Scurry”/ (already associated with her k-line for [Scurry]) through her p-sign { cul/di/} to a modified vocal expression similar enough to the conventional or common e-sign /”Scurry”/ to be recognized by others. In this general case, the infant’s p-sign functions as part of an indexical common sign (c-sign) because she accommodates her vocal expression to the conventional e-sign for the entity signified. In sum, every individual has p-signs as parts of k-lines associated with entities. When indexical signs are used in communication, it is because individuals negotiate the vocal expression of their signs to permit communication. In the special case above , we refer to these signs as s-signs; in the common case, we refer to these signs as c-signs. Both s-signs and c-signs are kinds of e-signs. The difference is that indexical s-signs are part of the infant’s idiolang. Indexical c-signs are part of the society’s public language. The need to associate infants’ idiolang-effective p-signs with others’ c-signs is a primary interior motor of the symbolic transition, as explained in the following.

The Theoretical Context of Modeling

We will model language development in the theoretical context of Minsky’s Society Theory of Mind and its suggested forms of representation. The central ideas used in setting the context of this model are that the general processing structure of the mind is represented by Minsky’s “ring closing” structures (p.205), with K-lines as the basic structural elements of memory (p.82 ff.). It is presumed that SR arcs grow into elaborate K-lines through the processes described below and may also grow into societies of agents, depending on the circumstances of learning. A first assumption is that one can think of the interior perceptions of “iconic” signs as Not-Significantly-Different from associated memories. A second is that incremental learning with respect to any SR arc proceeds in the Sussman paradigm, with B-brain structures modifying interrupted arcs, which leads to the recognition of e-signs for distinguished external things.

Indexical Signs and Learning Processes
The meaning of an iconic sign is determined entirely in the mind of the infant. The beginning of the indexical sign learning process is in the infant’s attempt to communicate something. Grant that the infant has associated some producible sounds, such as /cul/di/, with the entity [Scurry], through a listener agency mimicking what the infant perceives of the e-sign /”Scurry”/. {/Cul/di/} is then a p-sign which serves the infant’s expressive intentions. The e-sign/p-sign couple does not function as an indexical sign until there is negotiated a shared meaning between the infant’s talker agency and that of some other person who understands what the infant intends to communicate.
This can happen either through the infant improving her production of sounds to match better the conventional e-sign or by the other person changing his language to better communicate with the infant. In Peggy’s case, Bob started referring to [Scurry] by /cul/di/ when in her company. /Cul/di/ functioned as an indexical sign between Bob and Peggy because both used the same sounds to refer to the same thing signified by a convention, explicit here in Bob’s decision to adopt Peggy’s term for [Scurry]. This is a shared sign, or s-sign. Such signs are the main elements of the individual infant’s idiolang. Subsequently, Peggy modified the vocalization of her p-sign { /cul/di/} to conform to the sound /”Scurry”/ used by her mother and siblings when calling the dog or referring to her. This exemplifies the second process of negotiation of meaning between the infant and her local society, through which her idiolang p-signs are brought into correlation with the c-signs of the public language in her Society. Negotiating meaning is the common ground of these two different processes, even though the first is so transitory and the second is so dominant.
Negotiating meaning has been described as the primary means for learning to communicate with sounds. I argue that the portion of the process in the interior world — one linking modifiable responses to external signs (e-signs) in recognition processes with modifiable personal signs in expression processes — is the prototype for the symbolic use of words. Can we more precisely articulate this process to clarify the transition from using indexical signs to using symbols as a natural consequence of processes of communicating?

If the infant’s listener agency does not discriminate initially between what she perceives on hearing /”Scurry”/ and what her talker agency expresses as /cul/di/, there may be either improved discrimination with respect to /”Scurry”/ or better correlated expression by modifying /cul/di/ to become more like /skuh/ri/, or both. The first case is a modified discrimination on the recognition side of the SR arc. The second is one of modified production on the expression side of the SR arc.

We have already postulated that the infant’s p-sign generates a vocalization, but we have not previously noted that corresponding to the expressive aspect of the p-sign represented by the vocalization, there is a perceptual aspect which also needs to be represented. Let’s say that on hearing /”Scurry”/, the [Scurry] k-line perceives /skuh/ri/ but does not discriminate it as different from /cul/di/ through which it expresses p-sign {/cul/di/} . When subsequently a discrimination is made that perceived /skuh/ri/ is different from the normally produced /cul/di/ the infant’s listener and talker agencies have to negotiate an integration of the two aspects of the internal sign. It is this process that is the prototype of word definition by symbolic use of known words. Consider the following example:
We know that Scurry is a dog, and that /”dog”/ or /”doggie”/ are appropriate words to use in referring to [Scurry]. How could such a word enter the repertoire of indexical signs in Peggy’s idiolang ? We might say “Scurry is a kind of dog.” Peggy surely heard the word /”dog”/ before she had any well formed concept of [dog] as a category of kinds of entities. For her, when she understood that the term /”dog”/ was meant to indicate [Scurry], the sense she would have made through interpreting /”dog”/ would be that “dog” is another name for [Scurry]. /”dog”/ is thus a synonym for aspects of the p-sign represented either as { /skuh/ri/} or { /cul/di/} , depending on the time that the association was made.

Synonyms are common in an infant’s world. Mom and Dad have personal names. The infant herself may be Peg, Peggy, sweetie, or have any of a myriad of affectionate appellations. It is an open question as to the point in time at which anyone learns that what one assumes is a synonym for a well known entity’s name refers to some other entity. Recall, for example, that Peggy’s first two word sentence was /cul/di/va/va/! The context made it clear that she was referring to the barking of a remote dog. Did Peggy intend to communicate that ? Did she even know there were other dogs in the world besides Scurry ? No — this is a case of her marking no significant difference between her notion of [Scurry] and any other similar creature.
This reflection can help us understand one simple way that Peggy’s listener and talker agencies could integrate the perceived vocalization /skuh/ri/ of e-sign /”Scurry”/ with the expressive vocalization /cul/di/ of p-sign { /cul/di/} . Let it be the usual case to have multiple names for entities. Then /cul/di/ could be Peggy’s personal name for [Scurry] whom others may choose to refer to as either [/cul/di/] or /”Scurry”/. Some people might even refer to [Scurry] as /”dog”/ or /”chien”/. Development from the infant’s idiolang of indexical signs to a broadly flexibly public language then proceeds through elaboration of names for well known entities, where the relation of synonymity is the first wave of definition of new words (symbolic signs) in terms of preceding linguistic structures (indexical signs). It is in this specific sense that the integration of perceived and expressive vocalizations related to a specific k-line is the precursor and prototype of the development of symbolic thought, the key aspect of which is that words have meanings specifiable in terms of the network of meaning of other words.

.