The cognition of language and communication

G. Albyn Davis

For a long time, language has had a curious relationship to cognition in the vocabularies of rehabilitation practitioners, as well as laypersons. Diagnosticians have neatly divided and packaged disorders into separate categories. “Language” has been viewed descriptively with assistance from linguistics (e.g., phonology, morphology, syntax), whereas “cognition” has been identified broadly with “intelligence” and specifically with mental functions such as attention, perception, and memory. In some quarters, morphology and memory have been considered to be two separate entities, despite the reality that we store morphology in our memory. To assess memory, we use tests of cognition; to assess morphology, we use tests for aphasia.

It has been suggested that cognition plays a role in language and communication or that it is related to language and communication, as if “language” and “cognition” are different things. However, if cognition is identified with information processing and we think of language use as information processing, then it is consistent to think of language functions as embedded in cognition. Language comprehension and formulation are part of the cognitive system. When linguists characterize what we know about language, they are speaking of something in memory.

This first chapter introduces cognition and how it is studied, mainly in cognitive psychology. For the study of language processes, psycholinguists and many speech-language pathologists use the methods to be discussed. Then, the chapter provides an orientation to later topics of attention, memory, executive function, and language. Subsequent chapters will be more specific and expansive as to how cognition fuels language comprehension, formulation, and communication. Mainly, the present chapter sets up the thinking behind the investigation of cognition.

Assumptions in the study of cognition

Cognition is “an umbrella term for all higher mental processes . . . the collection of mental processes and activities used in perceiving, remembering, thinking, and understanding” (Ashcraft & Radvansky, 2010, p. 9). In contemplating their history, “cognitive psychologists generally agree that the birth of cognitive psychology should be listed as 1956” (Matlin, 2009, p. 7). Around this time, key publications and conferences steered psychology away from behaviorism. This change was driven by the Skinner-Chomsky debate over nurture versus nature, George Miller’s measure of short-term memory as being around seven units, and interest at Carnegie-Mellon University in the computer as an analogy for human information processing. The shift was complete when the Journal of Verbal Learning and Verbal Behavior became the Journal of Memory and Language in the early 1980s. Essentially, psychologists admitted that mental processes exist.

This section introduces three of four assumptions underlying the study of mental processes. They are presented as a hierarchy of dichotomies in Figure 1-1. First, there is a working distinction between behavior (as evidence) and what happens in our heads (as theory). Similarly for clinical diagnosis, we consider the relationship of what we can observe (symptoms) to what we cannot observe (diagnosed impairment). Scientists avoid writing statements like “comprehension is a behavior” so that they do not think carelessly and confuse one for the other.

Figure 1-1 A hierarchy of increasingly specific assumptions underlying the study of cognition.

Now that we are thinking inside the box, the second assumption differentiates the brain as a material thing from cognition as a mental thing. Because cognition is what the brain does and, therefore, is not truly independent of the brain, this dualism is largely a contrivance that is reflective of a research strategy. Cognitive psychologists approached their work as if “the mind can be studied independently from the brain” (Johnson-Laird, 1983). Through the 1980s, cognitive psychology texts barely mentioned the brain. At that time, Flanagan (1984) stated that cognitive psychologists “by and large, simply seem not to worry about the mind-brain problem.”

This dualistic approach was necessary, because technology for observing the brain (e.g., fuzzy structural imaging) was not matching the constructs for measurement of mental operations. Now, with the emerging fine-tuned technologies of functional neuroscience (Cabeza & Kingstone, 2006; Gazzaniga, Ivry, & Mangun, 2008), current editions of texts on cognition include chapters on the brain and sometimes are regaled with colorful pictures from brain imaging (e.g., Ashcraft & Radvansky, 2010). Nevertheless, one can conduct experimental cognitive psychology without considering the brain and, as a result, can restrict theory to functional matters (e.g., how memory works, as opposed to how the brain works).

In our everyday vocabularies, “brain” and “mind” often refer to the same thing. Yet, saying that someone has “lost his mind” does not mean that he has misplaced his brain (Box 1-1).

Box 1-1 Dialogue from a 1988 Episode of Miami Vice

Interrogator: What about the fact that he can’t remember any of his actions? Isn’t that a convenient lapse of memory?

Physician: The answer to your question is that I’m a neurosurgeon. You’re questioning Detective Crockett’s mental capacities. That determination should be made by a psychiatrist.

In his text for speech-language pathologists, Davis (2007a) encouraged clear thinking by recommending that we keep what happens to the brain (e.g., stroke, trauma) logically distinct from what happens to cognition (e.g., aphasia, amnesia). We can say that stroke causes aphasia (not that aphasia causes stroke). Neurosurgeons treat the brain, speech-language pathologists treat cognition, and so on. Whether cognitive-language therapy re-wires the brain is a current question. At least, to understand the nature of aphasia, we should have some idea of what happens to cognition.

Putting aside the brain, the third assumption focuses on cognition. Cognition consists of a fairly stable knowledge base and fleeting processes. This distinction was helpful when clinical pioneer Hildred Schuell proclaimed that what we do about aphasia depends on what we think aphasia is (Sies, 1974). A frequent question has been whether aphasia is an erasure of language knowledge or a disruption of language processing (while knowledge remains intact). The answer informs the broad approach to therapy, namely, whether it involves teaching words anew (because of a “loss” of knowledge) or exercising an impaired mental process that accesses a healthy store of vocabulary. Despite a layperson’s inclination to define aphasia as a “loss” of language, Schuell’s (1969) clinical experience led her to believe that “the language storage system is at least relatively intact” (p. 336). This belief, now supported by research, led her to advocate a “stimulation” approach to therapy.

Approaches to the study of cognition

Just as archaeologists build models of Troy based on analysis of unearthed floors and walls, cognitive scientists construct the most likely “functional architecture” of the mind from hundreds of carefully crafted experiments. Theoretical models are helpful in characterizing phenomena that are too big, too small, too old, or too obscure to be observed in everyday experience or “with the naked eye” (Davis, 1994). A layperson’s idea of “theory” can be heard in putdowns such as “it’s only a theory,” as if to say that such an explanation is only a guess. A scientific theory, however, is a collection of coordinated hypotheses built from appropriate evidence (Stanovich, 2007). Appropriate evidence of global climate change, for example, would be long-term worldwide temperature trends, as opposed to looking out the window (Box 1-2).

Box 1-2 What Is Appropriate Evidence?

Appropriate evidence is that which can be logically identified with or related to the mystery being studied. A National Geographic Network series about the science of migrations noted that we have not known what elephant seals do in the ocean for 10 months of the year. Our only data were based on watching them dive in. Speculation turned into scientific theory when data arrived in the form of tracking sensors attached to the seals as they swim, like electrodes on the skull to track neural activity hidden in the brain.

Clinical research sometimes entails collecting data and then exploring theoretical possibilities regarding the cause of observed behavior, called post hoc analysis. Theory-motivated clinical research, on the other hand, tends to lean on an established theory before an experiment is conducted. A useful theory, a priori, leads to a valid method and some predictions (i.e., “appropriate evidence”). The theory should be so clearly related to the experimental task that predictions of performance would logically and transparently follow from the theory. An investigator may think through what a person must do mentally to perform the task. Another approach, unfortunately, is to choose an established task created for other reasons (e.g., from a clinical test) without considering what a participant must do cognitively to carry out the task. A journal peer review may challenge an investigator to explain how a task demonstrates operation of the process purported to be studied. This disciplined and collaborative strategy maximizes the likelihood that a theoretical explanation is the correct one.

Any experiment consists of at least one comparison, either between groups of participants or between conditions. In clinical research, a study often contains both types of comparison. Differences between conditions are often labeled as special “effects,” such as the word frequency effect, the semantic priming effect, or the garden path effect. For example, a task with common words usually has fewer errors than a task with rare words. A theory of accessing the lexicon may predict this word frequency effect and then provide an explanation for why it does or does not occur.

A fundamental tenet is that a theory should be falsifiable, meaning that “in telling us what should happen, the theory must also imply that certain things will not happen. If these latter things do happen, then we have a clear signal that something is wrong with the theory” (Stanovich, 2007, p. 20). A common type of nonfalsifiable theory is one that is so general that it can explain anything (see Shuster, 2004). Explanations that are hard to test, such as appealing to motivations, are also difficult to falsify. Several comparisons should produce a pattern of results consistent with a theory, and the comparisons should also allow for the possibility of other patterns that could be suggestive of another theory. What follows are some of the basic approaches to making these comparisons in cognitive science.

A classic approach to experimentation is called mental chronometry or additive/subtractive methodology. Inspired by research to determine the speed of neural impulses, Franciscus Donders, a Dutch physician in the 1800s, used a subtraction method to measure the speed of mental operations in simple responses to lights. The experiment was a comparison between similar tasks. The general idea was that when one task takes longer, the difference in time is a measure of the operation that made the task take longer. In the 1960s, Sternberg (1975) worried that two tasks could differ in more than one way, spoiling theoretical interpretation. His solution, in order to study short-term memory scanning, included the comparison of several conditions differing in one respect (i.e., additive method).

While mental chronometry was paving the way for disciplined study of human participants, the computer metaphor for the human mind encouraged studies of computational modeling or simulation (based on “connectionist models”). Implying a comparison to human beings, research consists of “programming computers to model or mimic some aspects of human cognitive functioning” (Eysenck, 2006). Applying simulation to clinical populations, some investigators would artificially lesion a program to mimic a disorder (Dell & Kittredge, 2011). Wilshire (2008) made note of one strength in the simulation approach that is favored by cognitive theorists in general, namely, theoretical parsimony, or explaining the widest range of observations with the fewest assumptions.

A third approach is cognitive neuropsychology, in which brain-injured subjects are studied to test theories of cognition. Viewed broadly, this discipline can be any study of cognition involving brain-damaged people that has the goal of understanding normal function as well as dysfunction. In one branch of this field, “CN” focuses on single cases as examples of a lesion that hypothetically knocks out one component of a processing model underlying a simple task (Rapp, 2001; Whitworth, Webster, & Howard, 2005). A typical model of reading aloud given in Figure 1-2 provides a menu of possible impaired components. Although proponents of CN speak of testing theories of language in general, their research is restricted to single words. This narrow window on language and the absence of an automatic-controlled processing distinction were a concern for Wilshire (2008; also, Davis, 1989). She noted that computer simulation has an additional advantage of explaining how and the extent to which a cognitive component may be malfunctioning.

Figure 1-2 Typical cognitive neuropsychological model of stages in reading words aloud. Functional equivalents are noted in parentheses. [From Davis, G. A. (2007). *Aphasiology: Disorders and clinical practice* (2nd ed.). Boston: Allyn & Bacon/Longman.]

Attention

We must be aroused or alert (i.e., conscious) for intentional communication to occur and, once aroused, we establish awareness of our surroundings so that simple communication makes sense. This is the base level of attention. Then, when faced with multiple simultaneous stimuli in a conversation, we focus on something for in-depth processing. Inability to ignore irrelevant inputs can be an impediment to successful communication (i.e., the “cocktail party problem”).

“We use the term attention to describe a huge range of phenomena” (Ashcraft & Radvansky, 2010, p. 112). It is commonly considered to be a cognitive process that concentrates mental effort on an external stimulus or an internal representation or thought. Attending to external stimuli may be called “input attention,” which is the basic mechanism for selecting sensory information for cognitive processing. Input-directed attention includes an orienting reflex that directs us toward an unexpected stimulus and attention capture, which is driven by physical characteristics, namely, significance, novelty, and social cues.

Higher-level attention consists of different mechanisms. Selective attention (i.e., focusing) goes along with resisting distraction so that cognition becomes manageable. Another term, “spotlight attention,” is used for a focusing mechanism that prepares the processor to deal with information based on expectations. Selective attention is studied by presenting two stimuli and requiring response to one of them. Divided attention confronts multiple stimuli or processes at the same time. A dual task is used whereby a participant responds to two stimuli or performs two tasks simultaneously. A researcher is interested in the effect of dealing with one stimulus or task on the other. Discussions of attention mechanisms overlap with other aspects of cognition in that they contribute to resource allocation in working memory and the management of multiple tasks associated with executive function.

Memory

Ashcraft (1989) wrote that cognition is “the coordinated operation of active mental processes within a multicomponent memory system” (p. 39). A simple definition of memory is any retention of information in the mind beyond the life of an external stimulus (i.e., minimal memory). The ability to hold information in our head is fundamental to the mind’s (or brain’s) ability to perform even the simplest functions such as perception and recognition. Following the knowledge-process distinction mentioned earlier, the major components of the memory system consist of long-term memory (LTM) for passive storage of information and working memory (WM) for constraining the activity of processing.

Before these components are introduced, let us consider two questions that apply to both components of memory. How does information become represented in our heads, and what form does it take? This inner form is called a mental representation, which occurs either in permanent storage or in a transient state. A theory of neural representation can appeal to tissues and chemicals. Characterizing a memory in mental terms is more problematic. Resorting to analogy, we may think that a mental representation for a visual input might replicate the stimulus, like a photograph. An auditory stimulus may be replicated like a tape (or digital) recording. Testable hypotheses about mental representation are included in the collection of hypotheses comprising a theory of a language function.