Jerome A.Feldman
From Molecule to Metaphor
A Neural Theory of Language
Mit Press 2008


Feldman 135
Conceptual Schemas and Cultural Frames

The child's first words are labels for his or her experience, but not all experiences can be described with a single word. Actions such as „grasp“ or special relations such as „support“ inherently have multiple participants, or roles. Grasping requires roles for at least the grasper and the thing being grasped. Coordinated motor activities such as grasping are called motor schemas. The same term, schemas, is used to describe relational information as in the concept of „support“. Many of these cognitive structures are universal across all languages and cultures, and I refer to all of these as conceptual schemas or sometimes just schemas.

The embodied theory of meaning suggests that the child needs to have conceptual structures for understanding experiences before the words for labelling them can make sense. For example, every language has a notion of physical support, with roles for the supporter and the supported. The support schema is central to the literal meaning of the English word „on“ and to related literal and metaphorical meanings in other languages.

Other coherent collections of experience are particular to some culture, and following convention I refer to these as frames. Typically, a cultural frame such as the baseball frame will involve several basic conceptual schemas such as run, grasp, contact, and goal. Notice that a single word, like „shortstop“, can evoke the baseball frame with its many roles, actions, and relations.
In this chapter, I discuss both schemas and frames and also introduce notation for computationally modelling their roles.

Colour and the words that describe it are a particularly simple case of universal language tendencies. When cognitive scientists look for phenomena that are "universal", they cannot look at every language in the world, so they study several languages from different language families. If they find the same phenomenon in these unrelated languages, it is likely to be universal. Cognitive scientists have extensively explored the idea of conceptual schemas (like support) as universal, bodily based representations of experience.

Proposed universal schemas, such as support, arise from our common genetic heritage and shared developmental experiences. Every child learns how to perceive and understand quite a lot about his or her body and environment before learning language. No one knows how many universals schemas there might be, but best estimates are in the range of a few hundred.

For now, I focus on static schemas representing fixed relationships between things, such as the supporter and the object that is supported. In chapter 13, we extend the discussion to include actions, events, and other changes over time. In chapter 16, we examine how universals schemas and cultural frames form the metaphorical basis for abstract thought and language.

Image schemas

136 primitive image schemas
137 primary image schemas
138 universal image schemas
141 dynamic schemas

145 cultural frames
Image schemas are used to describe basic and universal packages of human knowledge. There are obviously also many packages of knowledge that are specific to a given culture, profession, or other entity. Charles Fillmore observed that to really understand the relationships between related words, you have to understand the structure underlying the conceptual setting which he called a frame. For instance, to understand the relationship between words such as „buy“ and „sell“ you have to understand the commercial event frame. In any frame, there are participants. Some frames also have scenarios.

146 Frames are important to our story in two ways. From the embodied perspective, we experience the world in coherent scenarios, and it makes sense that we should organise our knowledge this way. Words can be best understood as calling into mind (evoking) a frame. Many words can evoke the commercial event frame, including ones that are not explicitly mentioned in its definition including price, cost, and bargain.

There is also a more technical way in which frames are important for our purposes. Frames encourage us to describe events as a sequence of situations, each of which is described by a collection of features (like buyer) and fillers (like John). Both the idea of sequences of actions and their representation in terms of features and values are used throughout this book.

We can also consider frames for basic actions such as walking or touching a nearby object. When you think about it, we know remarkably little about how our bodies and brain actually carry out these actions. We saw the same thing earlier than discussing colour perception and division in general - we have no way to consciously Into the massive neural computation is that we know are responsible for our vision.

It is worth experiencing again this inaccessibility of detailed mechanisms. Try to figure out what is happening as you reach to touch some nearby object. In fact, this action involves elaborate coordination of many muscle and control systems including shifting your body poster to balance the movement. Although you probably didn't think of this, the motion is accompanied by adjustments in the visual system that take into account your changed head position and gaze.

So a great deal of bodily activity cannot be talked about directly and language or consciously thought about or modified. Of course, some features of our actions we can talk about and, it seems, these are the same features that we can consciously control. For example, in touching an object you can choose the body part that does the touching, the direction, speed, and force of your movement, and possibly a few other parameters.

147 Some indirectly relevant parameters include the properties of the target object and your reasons for doing the touching. And there are also some general control parameters that can accompany any action, for example, whether it is done once or repeatedly.

The limited accessibility of the details of actions raises an intriguing possibility: perhaps frames represent all we can say directly about bodily actions. This fits in beautifully with the embodied idea of word meaning. Each of us has rich experience with our bodily actions and perceptions, were we can say only a very restricted set of things about them. Since speakers and listeners share both the experience and the frame parameters, however, a word or expression can convey a great deal of meaning. The theory developed over the next few chapters relies heavily on the ideas that language primarily operates at the level of frame parameters and understanding involves imaginative simulation evoked by these frames.

This simple picture has two complications, and they are both important. While we have little direct control over the fine structure of a movement, there are many indirect ways of modifying our actions that can be readily conveyed in language. So you could reach to touch an object as if you were a baby, very old, very tired, frightened, and so on. This kind of language use also has an obvious space in experience (and mirror neurons). We discussed this in connection with metaphorical language, which plays a central role in the second half of this book.
The other complications with the idea of universal meaning frames is that human languages differ greatly in how they express things, including actions. Different languages have quite distinct ways of talking about the same underlying actions. The diversity of ways to describe actions across languages presents a challenge for frame semantics and, indeed, for any theory of language understanding. In fact, the packaging of information into a frames can differ markedly among languages, requiring a more fundamental cross-linguistic level of representation. Image schemas, including the primitive motor schemas, are the cognitive scientist‘s means of describing these language independent foundations of thought.

Image schemas and cultural frames figure prominently in the core semantics of human language. They form a crucial part of the computational level description of neural circuitry for language and thought.

171 If a common executing schema underlies action, recognition of action, planning, and simulation, then many questions and language become less mysterious. Shared and parameterized neural circuitry for executing and describing actions provides a natural grounding for semantics. We can understand someone by imagining ourselves in their situation, and we have apparently evolved mechanisms that do this automatically. Imagination or simulation will play a key role in our theory of embodied language understanding.


Feldman

HOME | | BOE