What is this article about?

This article states that the cognitive system has developed to facilitate action. Therefore, the article describes the functions of the cognitive system based on the implications for action. The author provides a model of cognitive structures and processes that underly the production of verbal and nonverbal behaviors during interactions.

First, they use a basic observation. This observation refers to that communicative behavior at first a novel and creative, but also patterned and repetitive. To explain, the social behavior of someone often consists of an assembly of repetitive words, topics, themes, and instrumental phrases. However, in every interaction, these elements are used in a unique combination. So, pre-established routines are adapted to immediate circumstances.

What is a model of the communicative output system?

To develop a model of the communicative output system, we have to keep in mind two basic processes of social behavior. These are the selection of old elements, and the construction of novel patterns.

These processes are related to memory storage, retrieval, and utilization. To explain this model, the authors use the approach of ‘cognitivism’.

What is ‘cognitivism’?

In cognitivism, theorists explain how the observed input-output takes place, by referring to the underlying processes and structures that are responsible for input-out processes. Also, in cognitivism, the structures and processes are theoretical entities. Thus, the structures are not always directly linkable to certain brain areas. These structures do have a link to brain areas, but this is not extremely important in cognitivism. Another fundamental aspect of this cognitive perspective is that in order to create predictions, a theory must specify the information structures of the mind and the processes (mechanisms) of these structures. If these properties are not specified, the model is not falsifiable and thus not scientifically correct.

Another important characteristic of cognitivism is that it can be used to create multiple, distinct models of which each can be correct. Thus, in cognitivism the goal is not to create ‘the perfect model’, instead the goal is to create a sufficient model.

Which model is described in this article?

Before diving into the theory, it is important to know that the human mind consists of both conceptual and procedural knowledge. Conceptual knowledge is knowledge about things, and procedural knowledge is knowledge about things that we have learned to do. The procedural knowledge consists of action-outcome contingencies (if we do this, then that happens). The elements of this memory are modular, which means that it is organized on the basis of a large number of elemental units. These elemental units are related to the output part. The action component consists of parameters that define complex behaviors.

These characteristics are important for the procedural memory. It is good that there is a large number of elements, because this means that the system is less complex. If the system was very complex, it would be difficult to produce novel behavior, because it would require a decomposition of complex elements. Also, it would require a lot of storage. Therefore, cognitive theorists concluded that people use “scripts” and “event schemata”. Another advantage of the fact that procedural information is stored as discrete units is that new information can easily be added. So, if a memory system consists of many modular elements, acquiring new knowledge is more easy.

Thus, the elements of the procedural system allow us to produce communicative behaviors. However, not all elements in this store are used during behaviors. This indicates that there is some process of ‘selection’. Before something in a system affects behavior, it must reach a certain threshold of activation. One way this happens is by past experiences that have led to desirable outcomes. If you know that something works well, you might activate this type of behavior. When you used an action in the past and it lead to a desired outcome, a procedural record is formed. Later, when you are in a similar situation, you can use this procedural record. Thus, this is one form of selection.

Another characteristic of the output system is that the output representation contains levels of abstraction. Also, there is a downward goal-setting influence. This influence leads to that abstract levels constrain the output of lower levels.

In sum, this section describes five different axioms (statements).

Axioms

• Axiom 1: A procedural record is a modular entity that contains a specification for action and an outcome associated with that action.
• Axiom 2: Each procedural record is characterized by a level of strength reflecting the status of the action-outcome contingency of the record. The strength of any record is a function of its recency, and frequency of activation.
• Axiom 3: The output representation of anticipated action is a hierarchy of levels of increasing specificity where each level is relatively autonomous in the execution of output demands.
• Axiom 4: At any moment, a procedural record has some level of activation. In order for a procedural record to impact output processing, the level of activation has to exceed some threshold value.
• Axiom 5: The activating conditions for any element of the procedural store are defined as occurrence of a goal to which that record is relevant, and the occurrence of any conditions which have proven to mediate the action-outcome contingency contained in the record.

What are the theoretical propositions?

In the section above, the characteristics of a cognitive model of the communicative output system were described. In this section, the goal is to provide a theory of the structures and processes of the output system.

Structural propositions

The authors describe 17 different propositions. First, you read elaborate information. Then, this information is summed in the proposition. So, first you have the information, and then the proposition related to this information.

Proposition  1

Interaction functions: These functions include goals that people try to achieve by communicating with others. There are three types of interaction goals: 1) specific situation-bound goals, 2) desire for a specific interpersonal relationship, and 3) management of personal identity.

1. Content formulation functions: These functions relate to the formulation of locutionary, illocutionary, and thematic dimensions of behavior.
2. Management functions: These functions relate to the need for topic continuity and chaining, and interaction initiation and termination which arise in the course of interaction.
3. Utterance formulation functions: These functions derive from lexical, syntactic, and articulatory requirements for the formulation of intelligent utterances.
4. Regulatory functions: These functions are related to speaker-turn regulation in ongoing interaction.
5. Homeostatic functions: These functions relate to the regulation of physiological controlled quantities during interaction.
6. Coordinative functions: These functions refer to the integration of effector units involved in articulation and nonverbal behavior.

It is expected that each socially competent adult has a number of procedural records that they use for each type of these functions. Each of these records relate to a specific outcome in a particular context. In other words: Procedural records are distinguished according to the nature of the outcome they represent. The outcomes that are relevant for the production of social behavior are the seven types of functions described above.

Another important cognitive structure for the production of communicative behavior refers to the output representation. This presentation if formed from the combination of activated procedural records. As described, this representation is hierarchic. Now, the authors describe four hierarchical levels of output representation.

“Procedural records can be distinguished according to the nature of the outcome they represent. The outcomes relevant to the production of social behavior are: 1) interaction functions, 2) content formulation functions, 3) management functions, 4) utterance formulation functions, 5) regulatory functions, 6) homeostatic functions, and 7) coordinative functions.”

Proposition 2

“The hierarchical levels of the output representation, respectively from abstract to specific, are: 1) Interactional representation, 2) Ideational representation, 3) Utterance representation, 4) Sensorimotor representation.”

Proposition 3

The interactional representation involves a representation of interaction related goals during social situations. So, it is an abstract representation of how interaction goals should be achieved. It includes a characterization of current condition, desired ends, and transition states that link the two. So, it is some kind of plan, that represents a number of states and transitions that lead to goal accomplishment.

“The interactional level represents an assessment of current state plus projected states and transitions leading to accomplishment of interaction goals.”

Proposition 4

The ideational representation is the second level of output representation. The interactional representation specifies a sequence that leads to goal accomplishment, and the ideational representation is used to implement a specific state change. Thus, the ideational representation specifies the content of a particular move or transition. For example, silent pauses during speech. The ideational representation arrives from a combination of procedural records that are relevant to content formulation functions and management functions. Thus, it involves the specification of: 1) propositional content, 2) illocutionary content, and 3) elements of thematic structure.

“The ideational level represents a single move, or transition, in the discourse and contains a specification of propositional content, illocutionary content, and elements of thematic structure”.

Proposition 5

The third level of output representation is the utterance representation. In this representation, syntactic, lexical, and phonological aspects of each utterance are formulated. The role of assembly processes in this level is to implement the content of the ideational representation in a linguistic string. The ideational representation is semantic, but the utterance representation is phonological. As noted, in the ideational representation, the production unit is a ‘move’ or a ‘transition’, and at the utterance level, the production unit is clausal in scope.

“The utterance level represents a single cause, contains a specification of lexical items and their order, and is phonological in nature.”

Proposition 6

The last level of output representation is the sensorimotor representation. In this level, commands are prepared. In the upper level of this representation, there are general motor schemata or programs. These are used to accomplish articulatory, regulatory, and homeostatic goals. In the lower levels, the current environment is analyzed.

“In cases of repeated assembly of a particular complex of procedural records, a unified assembly of those records may be stored in procedural memory.”

Proposition 7

The author states that next to procedural records and output representations, there are also other cognitive structures that are important in output processing.  He states that, just as procedural records may be inferred from past experiences, a small number of assembled representations may also be inferred from procedural memory.  These assembled representations are expected to develop through repeated assembly of a particular set of procedural records. Then, this assembly of procedural records function as one single element. At the interactional representation level, the assembled interactions may be available when routine activity results in accomplishment of desired ends. For example, the restaurant script: the script that people use when they go to a restaurant.

At the utterance level, the specific linguistic strings might be retained, so that no assembly is required at that level. Just as is the case in other elements of procedural memory, these assemblies of procedural records would be activated by a goal, and a set of relevant conditions.

“The activating conditions for a unitized assembly of procedural records are defined by the occurrence of a goal or goal complex to which the assembly is relevant plus any relevant initial conditions.”

Proposition 8

“An assembly of procedural records has a level of strength reflecting the status of the relations among its constituent elements. The strength of any such assembly is a function of its recency and frequency of activation, given that stored-action outcome contingencies continue to hold

Proposition 9, 10, and 11

The author also states that associative links can be developed between procedural records that are relevant to different levels of output representation. These links would then bias the selection of lower-level procedures by combining specific records to more abstract output representations. Then, a stored ideational assembly might be linked to a specific stored utterance. Thus, the move is always implemented via the same utterance. Also, these relations have reflexive properties, which means that activation of a low-level procedural record may lead to the activation of a higher-order record. So, a specific nonverbal behavior may lead to facilitate speech production by activation particular lexical or ideational items. For this proposal, the assumption is that repeated concurrent or serial activation of elements of the procedural store will establish an association between those elements such that activation of one will increase the activation level of the other.

• P9: “In addition to functionally-based relations between levels of the output representation, there may exist associative links between specific-procedural records of the various levels.”
• P10: “The strength of the associative links between levels of the output representation is a function of the recency and frequency of co-activation of the linked procedural elements.”
• P11: “The degree to which the activation of one procedural record impacts upon another is a function of: 1) the level of activation of the first record, and 2) the strength of the relation between the records.”

Processing propositions

To specific the nature of the activation process, three parameters are important: 1) level of activation, 2) the time required for activation, and 3) the demands on processing capacity required for activation.

Level of activation

Proposition 12

This proposition suggests that the most highly activated procedural records are those whose activating conditions are best fit for the current conditions.

“The level of activation of any procedural record is a function of the degree of match between current conditions and the activating conditions for the record.”

Time required for activation/speed of activation

It seems that, over repeated activation, the strength of a record will be increased.

Proposition 13

“The speed of activation of a procedural record is a positive function of the level of strength of the record.”

Processing demands

It seems that retrieval from procedural memory is automatic. So, it requires almost no attentional capacity.

Proposition 14

“The activation of procedural records occurs in parallel with no demands upon central processing capacity.”

However, even though this activation happens automatically, the integration of the activated productions does require processing capacity.

Proposition 15

“The assembly of activated procedural records requires central processing capacity.”

So, the assembly process does require cognitive resources. The factors that facilitate the formation of the output representation are then explained. First, if there are stored assemblies available, the assembly of a representational level is not required. Second, if a move has already been formulated, then no assembly is needed for the ideational level. This proposition is related to the well-practiced routines of people.

Another facilitation of output processing refers to that assembly of output representation in advance of the actual production reduces the load of assembly during interaction. This kind of preparation does reduce reaction times in cognitive and motor tasks. Also, planning in advance improves the fluency of speech, and reduces the amount of pauses.

Proposition 16

“Where stored assemblies or assembly of the output representation in advance of actual production obviates the need for assembly, social behaviors can be output with minimal demands on central processing capacity. The degree of processing capacity demands in such cases is a function of the amount of additional assembly which must be performed at the time of behavior.”

Lastly, the nature of the assembly process needs to be specified. The author states that assembly happens simultaneously across each level of the output representation within the limitations on capacity. Also, in each level, procedural records are serially incorporated, in order of their relative level of activation.

Proposition 17

“Within each level of the output representation assembly proceeds serially in order of level of activation of procedural records.”

What is the conclusion?

This theory uses different domains and describes two basic structures, and two basic processes. Because of the propositions, the theory is useful for describing the cognitive processes of communication.