The Way We Think: Conceptual Blending And The Mind's Hidden Complexities

In the fields of cognitive science and artificial intelligence, the most important unresolved problem concerns the methodology by which the human brain combines concepts from two or more disparate domains. The solution of this problem will result in a dramatic breakthrough in how we perceive human thought and how we can implement it in computing machines. To this day no scientist, technician, or engineer can claim to have a full view of how the human brain can take information about a musical composition for example and combine it with information detailing the rules of a chess game.It is easy to see that a musical composition can be thought of as a series of symbols written down in a definite order under the constraint of certain rules of harmony. A chess game may be viewed as a series of moves on a chessboard under the constraint of the rules of chess. These movies can easily be transcribed in symbolic form. Musical compositions and chess games can therefore be viewed as a series of symbols written on paper or computer screen. In this way one can view musical composition and chess playing as part of the same category, namely the category that describes how entities in time, here musical notes or chess moves, can be represented by symbols.But having this knowledge will not allow one to become a good composer and good chess player. For that one requires an understanding of the proper places in conceptual space that both of these activities lie, and how to combine these concepts in highly creative ways. What at first glance may be very disparate domains, musical composition and chess playing could be combined into one or more concepts that retain certain features of each domain but form a compact and effective knowledge base for which to compose music or play chess.The authors of this book call this process "conceptual blending" and have done an excellent job of presenting to the reader their research and commentary on how the human mind performs this activity. Many readers, especially those who demand their support from cognitive neuroscience, will of course view their opinions as controversial. But as a whole they are a good first step in trying to understand what might be called the combinatorics of concepts. Readers in the artificial intelligence community, especially those that are determined to implement these ideas in real computing machines, may find the book helpful but will no doubt also realize after reading it that much remains to be done for a total understanding of domain-general human thinking.Some of the examples that the authors use in the book are somewhat elementary but this was no doubt by design so as to make it more accessible to a general audience. There are places where this is not the case, as for example their use of the complex number system as being a "double-scoped network." Imagination plays a big role in conceptual blending, and this sets it apart from mere symbol manipulation, i.e. from the deceptions of the "Eliza" machine. They outline four different types of `blending' or `integration' networks and give real-world examples of each. As one might expect, the blending of concepts is a complex process, with the production of concepts never the result of applying just one mapping. It is also natural to expect that any logical inferences that take place in each of the domains may not survive when these domains are subjected to conceptual blending. The resulting blend may have vestiges (one might call them "shadows") of these inferences but the creative process involved in conceptual blending may result in inferences that are completely at odds with those in the original domains. Along these same lines it would seem that conceptual blending is irreversible, with this irreversibility even more apparent the more "entangled" the blended concepts are. This would raise the question as to the evolutionary advantage or energy requirements of conceptual blending, with the answer to this question no doubt arising from the view that a degree of compression is involved in it. Therefore the recollection/storage of huge knowledge bases becomes unnecessary, due to the ability to blend many ideas or concepts into a compact and useful form.A natural question to ask is whether these networks can be embedded in the computational paradigm, and if so, how computationally complex this implementation would be in real computing machines. The authors emphasize strongly that conceptual blending is not algorithmic, and so any machine implementation may require new computing paradigms and data structures than what have been developed hitherto. There are many researchers in the artificial intelligence community who are working feverishly to implement conceptual blending in some form or another. These efforts have been classified as "artificial general intelligence" and although most of this activity is outside of the realm of the academia, it has attracted the attention of many highly talented (and courageous) individuals.It might be tempting to view the theory of conceptual blending as outlined in this book as being one that could be easily viewed in terms of the field of category theory in mathematics. But in the latter concepts in one domain are related to another by "functors" that retain most the logic of each domain. Conceptual blending on the contrary mixes up this information and creates objects that could be very dissimilar to the ones in the starting domains. Of course, this does not mean that some sort of generalized category theory could not be invented that would emulate most of the features of conceptual blending. This would be an interesting research project for those who want to give the theory of conceptual blending a more rigorous mathematical foundation. In this regard a branch of mathematics called `topos theory' may be of assistance here.

I thought twice before taking two stars from this book because its content is so good. But I am evaluating the whole product so I had to consider the fact that this book has colored pictures which Kindle is unable to display and that are fundamental for the points being presented.I know I can go to Kindle PC and see the pictures there with all the colors etc. but one must agree that it is a pain in the butt to switch to the computer every time I needed to see a picture.So, if I had purchased the printed book, I would give it 5 stars, but the Kindle version has flaws.About the book's content, it carries very useful information if you look for something beyond a pure recipe book.

It's an interesting book that will challenge your thinking about you think and relate to different concepts in the world. At times, I think the authors stretch their theory, but overall I found this to be a solid book with good examples being used to illustrate what they meant by conceptual blending. I found the focus on language and metaphor to be particularly fascinating. With that said, this is a book best read slow and considered carefully. The concepts are complex and will challenge your thinking.

This book is brilliant and deeply thought-provoking. I read most of it when it first came out. Now, almost 10 years later, I bought it to deepen my understanding and use it in my research. It pays re-reading. It is even more brilliant than I knew - I can see that now that I have a greater depth of understanding.

I WANT TO SEE THE TABLE CONTENTS OF AMAZON BOOKS!!!

I think at some level this is a book that wants to live in two worlds, Academia, and the New York Times Bestseller List. But to live on the NYTBL, the text must understandable to the lay-reader, and this book is not "Cognitive Science for Dummies." Instead, the majority of this book is an exhaustive taxonomy of conceptual blending and its many parts in dry technical language. As an academic work, it may be brilliant, but I am not qualified to render that judgment. I only wished that while reading it, I had an instructor to go to for clarification. I did not have the background necessary to fully enjoy the intricacies of the subject matter. That said, I am still glad that I read it.Blending is the capacity to take two mental spaces, and connect them in certain ways such that a blended mental space emerges. What the reader finds in this book is that this sub-conscious mental facility is always at work, and that it is humans' advanced blending operations that in effect separate us from any other species on the planet. It is our heightened ability to blend that gave rise to art, science, and language.The best thing I took away from this read was a fascinating theory of the origin of language. It is well written and defended with rigorous logic.It is important to consider who should really read it though. It has potentially profound implications to the poet, the painter, the AI researcher, the philosopher, the teacher, and the parent, but I think one should also consider if they have the basis necessary to really "dig" what is being said here. I didn't, although I reiterate, I am glad I read it. So I guess the prerequisites are one three credit class in Cognitive Studies or Philosophy of Language. Alternatively, the neophyte could survive given the time and fortitude to do the research that will assist in making sense of this book as he goes along.Last note. If you do decide to read this one, make sure that you divine your own answer to the Buddhist Monk riddle before moving on to the next chapter, no matter how long it takes to achieve the answer. Doing this will really give you "global insight" into the difference between forms understanding and the development of a successful blend.

This is a good introduction and of course the original definitive book about Conceptual Blending sometimes called Conceptual Integration. Please remember that this stuff is still very very new having been produced in the last 10 to 15 years this book is 2002. For people who are dyed in the wool positivists (logical positivism a variant - look it up if you need to) this book seems woolly and vague because what they want is experimental method, empiricism, predictability and so on. The problem is fairly easily stated; those formal approaches have not borne fruit and yet billions have been spent and years of research undertaken yet even the most sophisticated robot can't do some of the simple things that a 4 year old does easily so what is going wrong?The way that we think is not the way computers do things. So how do we think, and that question is doggedly pursued without falling into that trap of positivist and formal approaches. If the formal approaches were so wonderful why don't we have robots already that operate and function as well as in the sci fi movies? The way people think is deceptively complicated much more than the average chess or GO playing computer.A lot of people are now working within the paradigm of Conceptual Blending and there is still lots to do so you do have to reserve judgement a little and let the book wash over you (take your time reading it) and ask the question, if not this, then what? Some of the examples in the book are a little involved, let us say but stick with it because you will find many examples that you can understand.Mark Turner has also produced another book in 2014, this time on his own about 'the origin of ideas' (it's called that) and perhaps this addresses some other questions about origins of input spaces within conceptual blending. I think this book How We Think is a milestone and was based on Koestler's work as well (and I probably have read everything that he ever wrote) particularly The Act of Creation which was part of a trilogy of books. How we Think is an Insightful, path breaking, brilliant and highly accessible work.

The theory of conceptual blending, as presented in this book, is one of those theories something like which has just got to be right. We start with notions familiar from AI and cognitive science such as frames and schemas, which provide the templates which the mind combines with concrete conceptual elements to generate representations of actual states of affairs in the world, known as spaces. The authors then demonstrate how a great deal of human thought, particularly those that cannot be captured with simplistic logic based semantics, such as imagination and creativity, can be described in terms of the blending of conceptual spaces. They then go on to show how thoroughly pervasive such blending is in constructing meanings. From as apparently simple as `John is the father of Jill', to the more richly evocative `beauty is in the eye of the beholder', on to the dazzling intricacies of say `from the ashes of disaster grow the roses of success'. They analyse the structures of such blending, how concepts and frames from one input space might be matched with elements in another input space, and the mechanisms whereby elements from each input are selectively merged into a new blended space. They describe four categories of blending, of which the most complex, the double scope blend, they argue is unique to humans, and is the cognitive capacity underlying all that is unique to humans. Their most provocative claim is that this capacity evolved about 50,000 years ago, based on what they conclude to be a relatively subtle adaptation in brain organisation, but from which exploded the uniquely human capacities of art, religion, culture, and above all, language. The theory presented is at least as explanatorily satisfying as any competing theories of language evolution. An aspect of the theory that I found particularly intriguing was the ubiquity of counterfactual processing in thought at every level, and its basis in blending. Also fascinating was the notion of character, as a central aspect of folk psychology, being deeply rooted in double scope blending.This has proved a hugely influential theory in cognitive science, and this book is its definitive manifesto. A global academic industry has emerged wherein conceptual blending theory has been applied to the analysis of a huge range of arts and humanities disciplines, from poetry to mathematics, from music to religion, and from psychology to linguistics. Judging from the range and scale of research it has initiated, it would seem there is a general sense that we have new tools to make sense of areas of human endeavour that we could hitherto only blindly grope for.I actually found that the analysis of examples that form the bulk of the book became a bit relentless, and ceased to make for such an enjoyable and stimulating reading experience as the initial presentation. I considered giving it four stars as such, but the ideas contained in the book are too important to be measured against my personal pleasure. For me the elephant in the room which the book does not begin to approach, and the most exciting aspect of the theory, is the computational basis of conceptual blending. As I understand it, we have got some way in getting neural nets to represent frames and concepts, for example see Feldman's  From Molecule to Metaphor: A Neural Theory of Language . But how might a network be made to support multiple simultaneously active conceptual sub-nets, and by what rules might they be made to constructively interact so as to generate novel blended concepts. I have to say that, given the decades it has taken to get as far as we have, the implications of conceptual blending make it appear, to me at least, that genuinely human-like machine intelligence is going to be even harder than we were able to imagine. It would thus seem that its realisation could well be postponed to a correspondingly remote future. From Molecule to Metaphor: A Neural Theory of Language

Ich habe das Buch vor drei Jahren gekauft und schaue seitdem immer wieder hinein. Ich habe es in zwei von mir gegebenen Seminaren zur Kulturtheorie verwendet (einmal nur peripher, einmal an zentraler Stelle).Es ist ein eindrucksvolles Buch; die Ansätze der "Mental Space Theory" und ihrer Weiterentwicklung "Blending theory" sind, hat man sie erst einmal verstanden, umfassend einsetzbar. Sie sind weit flexibler, und die entstehenden Analysen wirken damit weniger gezwungen, als die von Johnson/Lakoff und dem klassischen Metaphernmodell.Ich kann das Buch nur jedem dringlich empfehlen, der sich für die Möglichkeiten der kognitiven Semantik bzw. Kognitionswissenschaft interessiert. Auch wer die teils nervtötenden Selbstanpreisungen und ewiggleichen ideologischen Behauptungen ("embodiment" usw.) dieser Richtung längst satt hat, sollte es mit diesem Buch nochmal probieren, das viele Schwächen der früheren Ansätze nicht teilt (z.B. wesentlich präziser ist, nicht so viel Intuition/Beliebigkeit in den Analysen hat wie Johnson/Lakoff). Es geht zudem aus dem Bereich des Sprachlichen hinaus und versucht ansatzweise, "die Welt der menschengemachten Dinge" (Artefakte) zu erschließen, wobei natürlich nur einige Phänomene (Uhrenbeschriftungen; Messinstrumente etc.) erklärt werden.Wer Linguistik ist, die methodischen Grundlagen der Ansätze nachvollziehen oder selbst damit arbeiten möchte, sollte aber auf jeden Fall zusätzlich Fauconniers "Mappings in Thought and Language" (1997) ausleihen (gibt's wohl in jeder Unibibliothek). Da wird genauer erklärt, wie die Ansätze funktionieren. In diesem Buch werden die Analysen teils nur vereinfacht dargestellt, einige Fachterminologie wird einfach weggelassen. Zum Nachvollziehen reicht das aus. Große Stärke des Buchs ist die Breite der vorgeführten Beispiele und die Relevanz der Ergebnisse.

Headline says it all

No navigable contents page and index lacks page numbers, so the only way you can find something is if you know exactly what you are looking for. At only £2 cheaper than a printed copy, this exemplifies the disgrace that is ebook pricing.