Archive: September 9th, 2008

Lakoff and Johnson: Philosophy in the Flesh

[Readings] (09.09.08, 3:04 pm)


The Embodied Mind

The authors open the section by immediately making the connection to neural networks. They support this connection with (besides the obvious fact that our brains are made of neurons) evidence derived from cognitive science relating to how the perception motor areas of the brain interconnect. The neural argument is used to show how concepts and reason are embodied in nature. One of the mechanisms by which this occurs is categorization. Categorization is a quality of interaction with the world and is inherently embodied.

Category, concept, and experience are woven together inseparably. “An embodied concept is a neural structure that is actually part of, or makes use of, the sensorimotor system of our brains. Much of conceptual inference is, therefore, sensorimotor inference.” (p. 20) This claim is very philosophically charged, as it contradicts many of the accepted traditions of Western philosophy.

Note: An important thing to note about this, is that, if models are embodied, what does that mean for the capacity of games and software to communicate models? Arguments toward embodiment also support the importance of the emotional element in games and in fiction. Emotion is a visceral experience, which leads to a sort of world model feedback that is used in the mental processing of fiction (See Keith Oatley). Games and electronic media have a powerful capacity to represent models, but it is difficult to argue towards embodied cognition, but represent game worlds so abstractly. The irony in this is that simulated characters are represented as being embodied, but the actual human user lacks a thorough embodied experience with the simulation. There are ways of getting around that by arguing towards emotion and the success of similar works that are not heavily embodied, but it seems as though there should be something extra here.

Lakoff and Johnson give the label of “metaphysical realism” to the aspect of classical philosophy that asserts that the world is fully abstract and that it can be imagined and understood in a disembodied manner. Metaphysical realism asserts that our concepts reflect the world. The opposing argument is the idea of “embodied realism” which I would argue takes a more subtle approach: our concepts construct the world. The authors pose a model of perception developed by Berlin and Rosch, which consists of four conditions that define basic conceptual categories.

  • Condition 1: “The highest level at which a single mental image can represent the entire category.” Example is of a chair, table, car, etc. But, furniture does not fit into this category, as it is impossible to have a mental image of “furniture.”
  • Condition 2: “It is the highest level at which category members have similarly perceived overall shapes.” This has to do with recognition, and the ability to map a perceived object into the category.
  • Condition 3: “It is the highest level at which a person uses similar motor actions for interacting with category members.” This approach is an interactivity-based categorization. This idea is somewhat problematic, though, because it relates to the matter of affordances. I would argue that affordances work below a categorical level, but then, the goal of these conditions is to define the highest level that is intrinsically basic.
  • Condition 4: “It is the level at which most of our knowledge is organized. You have a lot of knowledge at the basic level. Think for a moment of all that you know about cars versus what you know about vehicles.” This is one of the more problematic aspects. The trend of AI pattern matching seems to operate at a level above the basic level, to higher level reasoning.

The essence of embodiment is that perception plays a central role in conception.

Primary Metaphor and Subjective Experience

This section resembles a great deal of the discussion in Metaphors We Live By. Here, the authors are establishing a type of metaphor that is used as a groundwork for explaining how metaphor is used as a fundamental building block for cognition. There are four parts to the integrated theory of primary metaphor:

  1. Johnson’s theory of conflation. Conflation is how somatic experience connects to foundational concepts during development. An example is the connection between warmth and affection experienced by infants. Conflation is paired with differentiation, wherein children separate domains, but the underlying association remains present.
  2. Grady’s theory of primary metaphor. These are like building blocks for larger metaphors: they are atomic metaphors, and primary units. Examples are simple associations such as “more is up”.
  3. Narayanan’s neural theory of metaphor. This theory uses the neural basis of cognition to explain how conflation is represented neurally via associations. This uses the neural groundwork of activation and association to explain how metaphors function.
  4. Fauconnier and Turner’s theory of conceptual blending. This idea suggests that when distinct conceptual domains are activated simultaneously, connections across the domains are formed, leading to new inferences.

There is a brief note in this section that is worthy of attention: “It is also important to stress that not all conceptual metaphors are manifested in the words of a language. Some are manifested in grammar, others in gesture, art, or ritual. These nonlinguistic metaphors may, however, be secondarily expressed through language and other symbolic means.” (p. 57) This idea conveys that metaphors operate beyond language, and extend into a much broader sense of meaning. If metaphor operates at the level of art or ritual, this seems to assert that metaphors are models at their essence, and perhaps all models are metaphorical systems. The claim is significant, but defensible, as models construct relationships between concepts within a domain, and metaphor is how relationships are constructed, ultimately mapping back to bodily experience.

The Anatomy of Complex Metaphor

On complex metaphors, the authors continue to stress that  models and metaphors are the same things: associative patterns.  Concepts are dependent on metaphors. Examples of complex metaphors are “A Purposeful Life is a Journey”, “Love is a Journey”, etc. These are explained as being tied together by various underlying sub-metaphors, woven together by associations which are experientially based.

Embodied Realism

The idea that has been at work in the past several sections is that realism is dependent on the body and on experience, and cannot be metaphysically known. In this section, the authors compare classical (or first generation) cognitive science with contemporary embodied cognitive science. First generation cognitive science is based on a priori philosophy, carrying into the study of mind all of the classical ideas from Cartesian philosophy. This approach prevented growth, by tying cognition down with philosophical commitments, and leaving it unable to answer experimental evidence. Second generation or contemporary cognitive science accepts the neural basis of cognition and as a result, its embodiment.

Reading Info:
Author/EditorLakoff, George; and Johnson, Mark
TitlePhilosophy in the Flesh
Tagsembodiment, metaphor, ai
LookupGoogle Scholar, Google Books, Amazon