Archive: October 10th, 2008

Michael Tomasello: The Cultural Origins of Human Cognition

[Readings] (10.10.08, 3:24 pm)


Tomasello’s problem is to understand how humans developed so rapidly in the evolutionary scheme. He suggests that a small genetic change enabled a process of cultural formation. Early primates can use tools, but they must learn individually. Humans have the power to build on top of existing knowledge. The underlying change that Tomasello argues consists primarily of intentionality, but also the qualities of imitation and identification.


A Puzzle and a Hypothesis

Tomasello is looking at the anthropological origin of human cognition. He is concerned with how cognition and complex behavior came so quickly the larger evolutionary scheme. He notes 3 categories of human development: Tools, Language (symbols), and Rituals.

“One reasonable hypothesis, then, is that the amazing suite of cognitive skills and products displayed by modern humans is the result of some sort of species-unique mode or modes of cultural transmission.” (p. 4)

The hypothesis provided is that humans have a species unique method of transmission of skills, rather than a biological one. The key element to this is a process that he calls the ratchet effect. Most animals use biologically inherited skills. Primates develop learned skills, but these are individually learned. Humans have the capacity to build skills over the course of development, where skills gained in one generation continue to the next.

Tomasello is looking at development, similarly to Vygotsky. Learning is dependent on others, and culturally mediated. To Tomasello, the key i all of this is identification. “These special powers come directly from the fact that as one human being is learning ‘through’ another, she identifies with that other person and his intentional and sometimes mental states.”

He argues for the uniqueness of human cognition because of traits at three levels. These traits are genetically based, but culturally implemented.

  • Phylogenetically (before birth), humans have the ability to identify with others.
  • Historiacally, development of artifacts and knowledge accumulates over time, and is does not start from scratch in each generation.
  • Ontogenetically (after birth), children develop in atmosphere of skills and signs, internalizing existing symbols and knowledge.

Biological and Cultural Learning

There is development in some animal species that is based on social interaction. This is called cultural transmission, and develops cultural traditions. This is a broad sense of tradition, though. The human difference in cultural transmission is in identification.

One of the differences between primate and other animal cognition is in the ability to recognize intentions. “Nonhuman primates are themselves intentional and causal beings, they just do not understand the world in intentional and causal terms.” (p. 19)

Later, intentionality is something that is attributable or projected onto other beings or individuals. Nonhuman primates fail in identifying causality: The example given on p. 22 shows a primates having a great deal of trouble with the trapped tube.

If we relate the understanding of intentionality as deriving from identification, we can see the influence of Vygotsky here. The essence of this is that “others intend because I intend.” This takes place throughout development, and varies with the developmental capability of the child. The gist is that when the child can have intentions, it can recognize intentions in others. This idea relates very closely to Lacan’s mirror stage, which occurs at around 6 months, where the infant begins to recognize itself. Initially, the infant is in opposition and at rivalry with its own image, but then comes to identify with it. This notion can be extended to identification with others.

Cases of nonhuman primate learning and culture: Tomasello attempts to debunk and critique the projection of social learning onto nonhuman primates. Characteristics of nonhuman primate learning:

  • Individual learning (not social)
  • Emulation learning (not imitative)
  • Ontogenetic ritualization (which is repeated responsive behaviors, not imitation)
  • No active teaching
  • Situational adaptation (not cultural development)

Human cultural development is intrinsically cumulative. Artifacts, which may be tools, rituals, or symbols, are developed between individuals, instead of within individuals. Thus, artifacts are gradually modified by each generation. This is the cultural ratchet. Imitation is necessary to pick up the existing base of a skill or artifact, and once that is imitated, then further development may occur. The process of ratcheting enables a history.

Another kind of ratcheting occurs between individuals through social interaction, and this is called sociogenesis. Tomasello looks explicitly at the subjects of language and mathematics.

Regarding language, Tomsaello argues that language is a gradual development: “The crucial point for current purposes is that all of the symbols and constructions of a given language are not invented at once, and once invented they often do not stay the same for very long.” (p. 42)

The idea suggested with this is that sociogenesis enables the construction of more and more complex ideas (citing the complex structure and function of languages). But, within communities, the essence of practice is to make complex ideas into simple ones. This can pull back to Lakoff’s notion of metaphor. To a developing individual, learning is about understanding complexities in embodied or familiar terms.

Regarding mathematics, in early civilization there were a large diversity of numeric representations. Eventually, Arabic numerals spread and became widely adopted. This suggests that the cultural ratchet operates on a very broad scale (across continents, even). The spread and adoption of ideas is also addressed by mimetics, where the idea is imitated and spreads. From the mimetic perspective, ideas that are good at being imitated (and utility positively affects this), will spread more readily. This idea is consistent with Tomasello’s emphasis on imitation, and emphasizes the notion that utility of ideas is not universal.

Tomasello makes a further distinction. Instead of the dichotomy of learned vs. innate, the dichotomy of ontogeny vs. phylogeny is more useful, and we should focus on ontogeny. What does that mean? What is the difference between it and phylogeny and learned behavior? In his description, ontogeny (at least in humans) has a special emphasis on imitation and intentionality. Ontogeny extends beyond learned behavior in that it is more than merely environmental adaptation or response. Ontogeny has to do with how behaviors emerge: “… the goal is not to decide whether some structure is or is not ‘innate,’ but rather to determine the process involved in its development.” (p. 51)

Joint Attention and Cultural Learning

There are 3 elements to early infant cognition: Understanding objects: Infants understand some principles behind how objects work, even before their capacity to manipulate them.

Understanding other persons: The have “built-in” facial recognition, and a capacity to imitate facial expressions. This is potentially a root of identification. Understanding self: They understand the limits of the self in manipulation, and will bail out of unachievable tasks. This is the stuff that appears during early development, and is similar to other primates.

At 9 months, a tremendous cognitive change begins to take place. This is manifested as a collection of behaviors that Tomasello calls joint attention. “But at around nine to twelve months of age a new set of behaviors begins to emerge that are not dyadic, like these early behaviors, but are triadic in the sense that they involve a coordination of their interactions with objects and people, resulting in a referential triangle of child, adult, and the object or event to which they share attention.” (p. 62)

Tomasello gives 3 accounts for the 9 month revolution, each of which is flawed in its own way.

  1. No strong cognitive changes take place as ability to interact is innate, and they possess some primary intersubjectivity, but infants lack motor capacity to express these interactions (Trevarthen 1979, 1993). This is countered by failure to reproduce results, and studies that reveal sophisticated motor skills.
  2. Infants are preprogrammed with capacity to interact socially, but this does not activate until the appropriate time (Baron-Cohen, 1995). The different social skills are separate and become activated one at a time. The data is inconsistent with this conclusion, though.
  3. The behavior around the 9-month phase is learned, and activated according to critical stimuli (Moore, 1996; Barresi and Moore, 1996). Again, observed data does not support this conclusion.

A suitable answer to this problem requires answers to the questions: Why do joint attention skills emerge together? Why does this happen at nine months?

The argument that Tomasello makes is that infants have an intrinsic ability to identify with others, and when the infant develops intentionality, then others may be understood as intentional agents as well. This is a projection of the self onto the other. The key element in this explanation is simulation. Infants may understand the other as like the self, simulating the other’s intentions in order to predict them. “Since other persons are ‘like me,’ any new understanding of my own functioning leads immediately to a new understanding of their functioning; I more or less simulate other persons’ psychological functioning by analogy to my own, which is most directly and intimately known to me.” (p. 71)

This example interrelates to the self centricity and absorbtion of toddlers. They use may their selves as a basis of understanding others, but cannot identify themselves as being beholden to the social conventions that others are subject to. This is somewhat at odds with Tomasello’s model.

The capacity to simulate is something that has been expressed by other cognitive scientists as important elements to cognition, but its development is not usually explained. For instance: Keith Oatley on interpretation of fiction.

The 9 month revolution occurs because at that point, the child becomes intentional (supported by Piaget), and it is able to identify that others are intentional as well.

Simulation is not an explicit conscious process, but rather an innate, embodied one. “My hypothesis is simply that children make the categorical judgment that others are ‘like me’ and so they should work like me as well.” (p. 75-76) Others are understood in an analogical relationship to the self. Intentional simulation is closely related to the construction of causal models, as relates to observations of physical phenomena. The intentional simulation hypothesis is supported by confirmed predictions with autistic children.

Behavior after 9 months has mimicry of intentional behavior, and further incorporation of intention to general engagement. “That is, whereas in early infancy there was some face-to-face dyadic mimicking of behavior, at nine months the infant begins to reproduce the adult’s intentional actions on outside objects.” (p. 81)

An interesting conflict occurs with playful behavior, which is construed as oppositional to intentional behavior. In play, intentional affordances are decoupled from the artifact. This is in conflict with Vygotsky, who suggest that playful artifacts are projections of unachievable desires. It seems that play would be a further example of projection and analogy, rather than decopuling.

Linguistic Communication and Symbolic Representation

Where did language come from? Symbolic representation is important because it is 1) intersubjective, and 2) perspectival. Language emerges from 1) joint attentional scenes, 2) communicative interaction, and 3) role-reversal imitation.

Language learning, and especially learning of the meaning of words comes from an identification of intentions within a joint attentional scene. “To acquire the conventional use of a linguistic symbol, the child must be able to determine the adult’s communicative intentions (the adult’s intentions toward her attention), and then engage in a process of role-reversal imitation in which she uses the new symbol toward the adult in the same way and for the same communicative purpose that the adult used it toward her.” (p. 117)

Joint attention is internalized into symbolic representation. This looks like the beginning of the internalization of social models or cultural identities. Objects are used as symbols, and this idea relates to the sense of pivoting. (p. 126)

Linguistic Constructions and Event Cognition

Children abstract from the concrete. They hear only concrete utterances, but are able to abstract them and understand linguistic structures from these. Tomasello explains that this process is very important for understanding how events are conceptualized. This idea goes back to models. Given concrete phenomena, children will develop models (intentional and causal) of how these phenomena work. There are inherent abstracting principles in model formation. The discussion here focuses on linguistic structures, suggesting that model formation depends on language.

Verbs are understood as embodied (kinematic and kinaesthetic) intentional experiences. Nouns are substitutable. This follows from joint attention: activity is intentional, but objects are targets for attention and may be interchanged. Here this level of structure is expressed in language.

Abstraction and schematization are the processes by which children form structures and categories in language. Concepts are formed and generalized (and overgeneralized) and later focused and refined. This relates back to Lakoff and Johnson. Concepts, models, linguistic constructions are developed, expanded, and used to match observed information. More interestingly, Tomassello hints (but does not address thoroughly) the idea of model divergence and refinement. This connects to conceptual blending, which explores the construction of new concepts from old ones.

Language is a tool for interpretation and conceptualization, that is, for forming and developing models. This supports the linguistic model of thought, and with intentionality, counters propositional models. However, Tomasello did hint earlier that models do occur before language. This suggests that humans have an inherent power for using models, but it is though language that these models can be most readily changed and manipulated.

Discourse and Representational Redescription

“The current hypothesis is that the perspectival nature of linguistic symbols, and the use of linguistic symbols in discourse interactions in which different perspectives are explicitly contrasted and shared, provide the raw material out of which the children of all cultures construct the flexible and multi-perspecitval–perhaps even dialogical–cognitive representations that give human cogniition much of its awesome and unique power.” (p. 163)

The interesting element here is the multi-perspectival nature of using language. This echoes back to the issue of identification. Dialogical cognitive representations seems related to the idea of simulation. Tomasello is saying here that cognition is powerful because of multi-perspectival ability.

The function of discourse: negotiating the form of an utterance from its content. This adopts a symbolic view, but one that is not propositional. Tomasello argues for a model based view of communication. The feedback in discourse enables feedback on model construction. Reconciling differences relates to synthesizing and blending models. (p. 171)

There is a question posed here: Intentional agency versus mental/belief/moral agency. He contrasts between theory-theory and simulation theory. Both are rationalizations for how children understand others as having varying beliefs. Tomasello explains the understanding of varying beliefs as a natural and gradual consequence of development. (p. 174)

Reading Info:
Author/EditorTomasello, Michael
TitleThe Cultural Origins of Human Cognition
Tagsanthropology, linguistics, specials
LookupGoogle Scholar, Google Books, Amazon