JB: Where are you headed next?
LAKOFF: I've plunged myself as fully as possible into the research that Jerry Feldman and I have been doing for the past decade at the International Computer Science Institute on the Neural Theory of Language (www.ics i.berkeley.edu/NTL). That's where most of my technical research effort is going to go for quite a while.
Jerry developed the theory of structured connectionism (not PDP connectionism) beginning in the 1970's. Structured connectionism allows us to constructed detailed computational neural models of conceptual and linguistic structures and of the learning of such structures.
Since 1988, we've been running a project takes up a question that has absorbed both of us: From the perspective of neural computation, a human brain consists of a very large number of neurons connected up in specific ways with certain computational properties. How is it possible to get the details of human concepts, the forms of human reason, and the range of human languages out of a lot of neurons connected up as they are in our brains? How do you get thought and language out of neurons? That is the question we are trying to answer in our lab through the computational neural modeling of thought and language.
JB: How do you connect structures in the brain to ideas of space?
LAKOFF: Terry Regier has taken the first step to figuring that out in his book The Human Semantic Potential. He has hypothesized that certain types of brain structures - topographic maps of the visual field, orientation-sensitive cells, and so on - can compute the primitive spatial relations (called "image-schemas") that linguists have discovered. The amazing thing to me is that not only do we actually have a reasonable idea of how certain types of neural structures can give rise to spatial relations concepts. Recent neural modeling research by Narayanan has similarly given us an idea of how brain structures can compute aspectual concepts (which structure events), conceptual metaphors, mental spaces, blended spaces, and other basics of human conceptual systems. The next breakthrough, I think, will be a neural theory of grammar.
These are remarkable technical results. When you put them together with other results about the embodiment of mind coming from neuroscience, psychology, and cognitive linguistics, they tell us a great deal about things that are important in the everyday lives of ordinary people - things that philosophers have speculated about for over 2500 years. Cognitive science has important things to tell us about our understanding of time, events, causation, and so on.
JB: Like what?
When Mark Johnson and I looked over these results from the cognitive sciences in detail, we realized that there were three major results that were inconsistent with almost all of Western philosophy (except for Merleau-Ponty and Dewey), namely:
The mind is inherently embodied.
Most thought is unconscious.
Abstract concepts are largely metaphorical.
This realization led us to ask the following question in Philosophy In The Flesh: What would happen if we started with the new results about the mind and reconstructed philosophy from there? What would philosophy look like?
It turns out that it looks entirely different from virtually all the philosophy that went before. And the differences are differences that matter in your life. Starting with results from cognitive semantics, we discovered a lot that is new about the nature of moral systems, about the ways that we conceptualize the internal structure of the Self, even about the nature of truth.
JB: This seems like a distinctively new kind of enterprise.
LAKOFF: It's an interesting enterprise to take philosophy as a subject matter for empirical study in cognitive science. Most philosophers take philosophy as an apriori discipline, where no empirical study of the mind, reason, and language is necessary. In the Anglo-American tradition, you are taught to think like a philosopher and then it is assumed that you can, on the basis of your philosophical training, make pronouncements about any other discipline. Thus, there are branches of philosophy like the Philosophy of Language, the Philosophy of Mind, the Philosophy of Mathematics, and so on. Johnson and I realized that philosophy itself, which consists of systems of thought, needed to be studied from the perspective of the cognitive sciences, especially cognitive semantics, which studies systems of thought empirically. Our goal has been to bring a scientific perspective to philosophy, especially a perspective from the science of mind.
JB: How does this connect with traditional philosophy?
Lakoff: It is a startling thing to realize that most of Western philosophy is inconsistent with fundamental results from the science of the mind. But that is negative. We respect and value philosophy. Our work comes out of a deep love for philosophy and a disappointment over what it has been over the past couple of decades. We wanted to look at great moments in the history of philosophy - the Presocratics, Plato, Aristotle, Descartes, Kant - even the analytic philosophers - and show what shining the light of cognitive science oh philosophy could reveal about the nature of philosophy.
What we discovered was fascinating: Each major philosopher seems to take a small number of metaphors as eternal and self-evident truths and then, with rigorous logic and total systematicity, follows out the entailments of those metaphors to their conclusions wherever they lead. They lead to some pretty strange places. Plato's metaphors entail that philosophers should govern the state. Aristotle's metaphors entail that there are four causes and that there cannot be a vacuum. Descartes' metaphors entail that the mind is completely disembodied and that all thought is conscious. Kant's metaphors lead to the conclusions that there is a universal reason and that it dictates universal moral laws. These and other positions taken by those philosophers are not random opinions. They are consequences of taking commonplace metaphors as truths and systematically working out the consequences.
JB: What's the import of recognizing that metaphors are central to the work of earlier philosophers?
LAKOFF: It is not just earlier philosophers, but contemporary philosophers as well. Our moral is not that their work should be disregarded because it is metaphorical. Quite the opposite. Because most abstract thought is, and has to be, metaphorical, all rigorous abstract systems of thought will be like those of the great philosophers whose systems of thought we analyze. Moreover, everyone's everyday reasoning is often of the same character, though hardly as consistent overall. A cognitive perspective on philosophy not only teaches us how the great philosophers thought, but it gives us deep insights on how all of us think - at least when we're being consistent and systematic. It also tells us that, in most cases, the answers to the deepest questions of human existence will most likely be metaphorical answers. There is nothing wrong with this. We just need to be aware of just what our metaphors are and what they entail.
Another positive thing we sought to do was to look at the most fundamental of philosophical concepts from the perspective of cognitive semantics. Mark made a list of the basics. In addition to Truth, we looked in detail at Time, Causation, Events, The Mind, The Self, Morality and Being. Luckily, a fair amount of work had already been done on these within cognitive semantics. We pulled the results together, unified them, and worked out further details. Not surprisingly all of these abstract concepts turned out to be mostly metaphorical, using multiple metaphors, each with a different logic. Thus, there is not one concept of causation, but around 20, each metaphorical and each with different inference patterns. Thus, causes can be links, paths, sources, forces, correlations, essences, and so on. Pick a metaphor for causation and different inferences come with the metaphor.
The science and the social sciences all use causal theories, but the metaphors for causation can vary widely and thus so can the kinds of causal inferences you can draw. Again, there is nothing wrong with this. You just have to realize that causation is not just one thing. There are many kinds of modes of causation, each with different logical inferences, that physical, social, and cognitive scientists attribute to reality using different metaphors for causation. Again, it is important to know which metaphor for causation you are using. Science cannot be done without metaphors of all sorts, starting with a choice of metaphors for causation. Most interestingly, if you look at the history of philosophy, you will find a considerable number of "theories of causation." When we looked closely at the philosophical theories of causation over the centuries, they all turned out to be one or another of our commonplace metaphors for causation. What philosophers have done is to pick their favorite metaphor for causation and put it forth as an eternal truth.
JB: Where does morality come into all this?
LAKOFF: One of the most satisfying set of results is the collection of metaphors governing moral thought. We found that they all seem to arise naturally in an embodied way from forms of well being - health, wealth, uprightness, light, wholeness, cleanliness, and so on. A particularly interesting result is that moral systems as a whole seem to organized metaphorically around alternative models of the family. Again, this should not be surprising, since it is in our families that we learn what we take as moral behavior.
We are now in a position to study the metaphorical structure of various moral systems. We think that cognitive science allows one to give much more detailed and insightful analyses of metaphorical systems than has ever been available before. For example, in our study of Kant's moral theory, we argue that this great intellectual edifice arose from just four basic metaphors, and that this allows us to see just how the various aspects of Kant's moral theory fits together.
Cognitive Science not only sheds light on the conceptual structure of moral systems, but also on politics and social issues. Some colleagues and I are now in the process of forming a political think tank to apply these methods of cognitive analysis to everyday political and social issues.
Perhaps the most sobering result is the most fundamental. We are neural beings. Our brains take their input from the rest of out bodies. What our bodies are like and how they function in the world thus structures the very concepts we can use to think. We cannot think just anything - only what our embodied brains permit.
Metaphor appears to be a neural mechanism that allows us to adapt the neural systems used in sensory-motor activity to create forms of abstract reason. If this is correct, as it seems to be, our sensory-motor systems thus limit the abstract reasoning that we can perform. Anything we can think or understand is shaped by, made possible by, and limited by our bodies, brains, and our embodied interactions in the world. This is what we have to theorize with. Is it adequate to understand the world scientifically?
There is reason to think that our embodied conceptual resources may not be adequate to all the tasks of science. We take case studies from physics and discuss them in our sections on Time and Causation. General relativity is a good example.
JB: So, what's the big change here?
LAKOFF: In characterizing space-time, Einstein, like Newton before him, used the common metaphor that time is a spatial dimension. My present time and location is metaphorically conceptualized as a point in a four-dimensional space, with the present as a point on the time axis. In order for there to be curvature in space time, the time axis must be extended - it cannot be just one point, the present. In addition to the present, the time axis must include portions of the time axis understood as future and past if there is to be enough of the time axis to form a curved space time. This seems to imply, as philosophers have repeatedly observed, that at least portions of the future and past coexist with present. And if the future exists at present, then the universe is deterministic. Frankly, it seems nutty to say that the past, present, and future are coexistent - and yet the curvature of space-time seems to imply it.
JB : Does the problem lie with the physical theory or the mathematics used to express it?
LAKOFF: It lies with the common metaphor "Time Is A Spatial Dimension", which is used to understand Einstein's mathematical theory of the physical universe. The philosophical entailment of determinism is coming not out of the mathematical physics, but out of that metaphor applied to the mathematical physics. Does that mean that we should-or can-try to jettison the metaphor?
For better or worse, we cannot get rid of it - even if it does have a nutty entailment. Physics is about something. We need to link the mathematics of relativity to an understanding of space and time. "The Time Is A Spatial Dimension metaphor does that job." We have no better metaphor and no literal concept arising from our embodied minds to replace it with. The commonplace metaphor may be imperfect in having a nutty entailment, but it's the best that embodied human conceptual systems are likely to come up with. What this means is that it is important to separate the mathematical physics from the commonplace metaphors used to comprehend it. And it is vitally important not to take those metaphors literally, even if that leaves us with no literal understanding at all. We should not take time literally to be a spatial dimension; we should recognize that we are using a common metaphor, and that the metaphor has the unwanted baggage of determinism-the entailment that present, past, and future coexist.
The moral is that you cannot take conceptual systems for granted. They are neither transparent nor simple nor fully literal. From the perspective of the science of mind, science itself looks very different from what we are commonly taught it is. Scientific understanding, like all human understanding, must make use of a conceptual system shaped by our brains and bodies.