Defining thinking is inevitably difficult because thinking is already implied behind the action of defining. The task of understanding thinking is a bit like the short cartoon above in which the Pink Panther sucks himself up entirely in a vacuum cleaner that he himself is holding – this sucking yourself up in your own vacuum cleaner move is not really possible but we enjoy imagining it as if it was possible.
 
When people write and talk about ‘teaching thinking’ they do not just mean teaching any and all types of thinking because some thinking is obviously quite bad. They mean teaching ‘good thinking’ which they might call ‘intelligence’ or ‘higher order thinking’ or some other technical sounding term which always really translates most accurately as ‘the kind of thinking that we do not think we see enough of and that we want to see more of’. Of course this kind of thinking changes over time. It is plausible that Plato and other ancient Greeks valued formal logical reasoning so much precisely because there was not much of it around and it was very hard for humans to do. Now there are computers who can do that sort of thing for us we tend to value creative reasoning more. In 1987 Lauren Resnick reported that the kind of thinking that teachers of thinking valued was characterised by complexity, nuance, multiplicity, uncertainty and surprise by which they meant that they could not say for sure in advance what good thinking was but they recognised it when they saw it (Resnick, 1987).
 
In order to try to account for the nature and origin of the complex and surprising thinking that we value and want to teach more of, dialogic theory puts forward the metaphor of thinking as embodied dialogue. Educational research has confirmed that the metaphor of thinking as dialogue is a fruitful one. This metaphor lies behind programmes that have been successful at teaching thinking (Resnick et al, 2015). Bakhtin’s clarification that dialogue occurs only when answers give rise to new questions (Bakhtin, 1986) is useful and suggests that the ‘dialogic’ nature of dialogue refers to dynamic mutual illumination and inter-animation across a gap of difference (Wegerif, 2011).
 
Considering what is wrong with some of the other metaphors for thinking provided by cognitive psychology can help us to understand the potential value of the metaphor of thinking as dialogue. Behaviourism gave us a metaphor of thinking as ‘nothing but talking to ourselves’ producing sub-vocalisations that can be measured as behaviour (Watson, 1958), cognitivism violently replaced behaviourism with the metaphor of thinking as nothing but information processing (Pinker, 1999) and now I hear it is hard to get a job in a psychology department unless you can scan brains and the new metaphor of thinking is nothing but neural activity. These various metaphors or lens have all proved insightful in different ways. The problem with them arises from their ‘nothing but’ nature. Dialogic theory says that the ‘truth’ of something complex like ‘thought’ is unlikely to be found in any one metaphor or model but is something that we can best move towards when we have a variety of different metaphors in dialogue with each other. This is the polyphonic version of truth put forward by Bakhtin (Bakhtin, 1991). It also, interestingly, lies behind the new transdisciplinary understanding of how best to make progress in science (Nowotny, 2004).
 
Dialogue as a metaphor for thinking is both specific and general. Specifically it draws attention to and promotes real concrete face to face dialogues in classrooms when the thinking is found in the speaking or other embodied semiosis. The way that small groups work together to solve problems and to pose problems is already thinking (Woolley et al 2010: Stahl, 2009). The way in which cultures, societies and communities respond to challenges and design together for a collective future is also thinking (cf Dewey on the importance of ‘social intelligence’ 1993, p104).
 


The word ‘teach’ comes from an old German root ‘tǣcan’ which means to point out. Teaching is about pointing things out. It is only possible to point things out in the context of a relationship where you can follow my gaze and I can follow yours. So teaching involves first building a relationship and then directing the attention of students within the context of that relationship.

If we want to teach thinking then the most important thing to point out is ignorance: just how much we do not know. Rather than simply being a model for ignorance, by asking questions a teacher can serve as a model for curiosity.
 
Teaching questioning 

People might say that questions are always asked in a context and are always questions about something. I am not so sure. I think that, independent of any language or any sign system there is somehow always the archetypal question. Not a questioning of this or that but just a general questioning. An attitude of curiosity about new things and humility in the face of the many things that we do not know – especially the incalculable number of the things that we do not even yet  know that we do not know.
 
Whatever you point out to learners it is possible to point it out in a way that closes down this fundamental question or in a way that opens it up. Knowledge, as this is taught in schools, is only the dialogue so far. I mean by this that is consists only of answers that have been given to questions that have been put. Teaching knowledge not as finished and final but as the story of this dialogue leaves a space for the learners to enter into knowledge as an ongoing dialogue in which they themselves are able to ask further questions and find further answers. In this way anything and everything can be taught as an invitation to join a dialogue and so as an invitation to think (Langer, 2016).
 
Opening, widening and deepening dialogic space 

Constructivist accounts of thinking tend to emphasise the positive ability to build models and systematically apply thinking tools (Holyoak and Morrison, 2005). A dialogic approach lays more stress on what the poet Keat’s referred to as ‘negative capability’ or the ability to remain in uncertainty until a creative solution emerges (Keats, 1817).
 
Simply pausing after asking a question is a good illustration of what it might mean in practice to teach thinking by opening a space.

The kind of talk moves promoted in dialogic education usually include asking open questions such as ‘why do you think that?’. Such moves do not work as positive tools to co-construct meaning but as a negative and indirect way to open a space for reflection and the resonance of multiple voices out of which a creative response might (or might not) emerge.
 
Dialogic switch in perspective

In a dialogue we sometimes do not understand the other person’s point of view initially and have to listen harder which means doing work to re-construct it so that it makes sense and we can practice inhabiting it ourselves. This switch in perspective to facilitate understanding is not a once and for all switch, we do not lose our initial perspective in making the switch, but it is more about being able to hold different perspectives in tension together. The ease with which children can make this switch depends on the quality of their relationships. However nicely children talk together to ask each other questions and give each other reasons this will not automatically translate into insight unless they allow themselves to switch positions with other speakers. Such switches do not only occur with physically present voices and physically present tools but also with virtual cultural voices, voices such as Shakespeare, History or Mathematics (e.g Kazak, Wegerif, & Fujita, 2015).
 
Identification with dialogue

Different ways of talking in classrooms are related to different kinds of identification (Wegerif and Mercer, 1997). Where children identify with themselves only and reject the other they might be prone to what Mercer calls ‘disputational’ talk (Mercer and Littleton, 2007). However when they identify strongly with their group they might be prone to what mercer calls ‘cumulative talk’ and what is often referred to in psychology as ‘group think’ which is when the harmony of the group prevents critical questioning and good reasoning. Issues of identification seem important to group thinking and one mechanism of successful dialogic thinking might be shifting that identification away from all static bounded objects, be that an image of the self or an image of the group, onto identification with the open-ended process of dialogue itself. (Wegerif, 2011).
 
Changing the culture

Some of the proposed mechanisms or processes for understanding why dialogic education works to promote better thinking focus only on changes within individuals. But individuals are shaped within cultures. One way to understand this, informed by Rom Harre’s positioning theory (Harre, 1999), is about how different cultural ‘discourses’ offer different ‘speaker positions’. In standard classroom cultures, for example, students are often positioned as not being able to initiate dialogues. An element that is common to all dialogic education approaches is a concern to address behavioural norms directly by explicitly questioning old norms and teaching new norms or what Mercer calls ‘ground rules’. These new ground rules or behavioural expectation in turn shape how individuals see themselves and their possibilities. Teaching ground rules is a way of teaching thinking through changing the culture such that a different experience of individual agency is produced, an experience that is less egotistical, less tribal and more tolerant of uncertainty and multiplicity because open to learning from the others and from otherness.
 
Design for collective global intelligence

Technology is part of the essence both of education and of thinking. We know that the dialogues of oral thinkers like Socrates, Gautama Buddha and Confucius were intelligent only because their followers wrote them down. Because of the technology of literacy and because of mass education policies we have something that Oakeshott referred to as the ‘Conversation of Mankind’ (Oakeshott, 1962). Education, as we know it today, is part of the communications technology that enables us to participate in this great dialogue of culture, - without education there is no literacy, mathematics or science. The advent of the Internet is transforming this ongoing dialogue of humanity, bringing it into real-time. Through access to the Internet we can all potentially participate in global dialogues that are building shared understanding and knowledge in every area as well as responding to challenges and designing the future together. Realising this new potential for real-time global collective intelligence requires teaching thinking as a form of educational design. Some argue that our use of the Internet is leading to increased stupidity and tribalism. But it is possible that the use of the Internet, with the right kind of supporting education into how to dialogue across difference more constructively, might be the beginning of a new age of collective global intelligence.
 
Summary

The metaphor of thinking as dialogue leads to an understanding of teaching thinking as drawing students into dialogue. This has some overlap with other models of teaching thinking and also some differences. It overlaps in seeking to produce thinking dispositions such as curiosity and thinking strategies such as asking questions and reasoning. The main differences stem from the understanding that dialogue between people and dialogues carried by media in society as a whole is already thinking such that silent inner thought is just a modality of this larger dialogue. One aspect of teaching thinking is to be concerned to teach individuals to think through internalising dialogue such that they end up carrying their own inner dialogic space around with them. But more than that a dialogic approach to teaching thinking is concerned to open, widen and deepen shared spaces of dialogue in the school classroom and beyond. A dialogic theory of teaching thinking suggests that it is important to teach cultures to think as well as individuals and ultimately to teach our increasingly global society to think. The project of teaching thinking through engaging students in dialogue therefore connects a focus on dialogues in classrooms to the design of educational technologies, including pedagogies, which will promote and sustain a more intelligent global dialogue.
 
 
 
References

• Bakhtin, M. (1981). Discourse in the novel. In M. M. Bakhtin (Ed.), The dialogic Imagination. Four essays by M.M. Bakhtin. Austin: University of Texas Press.
• Bakhtin, M. M. (1986). Speech genres and other late essays. Austin: University of Texas.
• Dewey, J. (1993). The political writings. Hackett Publishing.
• Harré, R. (1999). Positioning theory. The International Encyclopedia of Language and Social Interaction. New York, NY: John Wiley and Sons
• Holyoak, K. J., & Morrison, R. G. (20015) Introduction in (Holyoak, K. J., & Morrison, R. G Eds.). The Cambridge handbook of thinking and reasoning. Cambridge University Press.
• Kazak, S., Wegerif, R., & Fujita, T. (2015). The importance of dialogic processes to conceptual development in mathematics. Educational Studies in Mathematics, 90(2), 105-120.
• Keats, J. (1817). Letter to George and Thomas Keats. December, 21(27), 1817.[ https://en.wikisource.org/wiki/Letter_to_George_and_Thomas_Keats,_December_28,_1817]
• Langer, E. J. (2016). The power of mindful learning. Hachette UK.
• Mercer, N. and Littleton, K. (2007). Dialogue and the Development of Children’s Thinking: A Sociocultural Approach. London: Routledge
• Nowotny, H. (2004). The potential of transdisciplinarity. H. Dunin-Woyseth, H. and M. Nielsen, Discussing Transdisciplinarity: Making Professions and the New Mode of Knowledge Production, the Nordic Reader, Oslo School of Architecture, Oslo, Norway, 10-19.
• Oakeshott, M. (1962). The Voice of Poetry in the Conversation of Mankind, Rationalism in Politics and Other Essays. London: Methuen: 197–247.
• Resnick, L., Asterhan, C., & Clarke, S. (2015). Socializing intelligence through academic talk and dialogue. American Educational Research Association
• Stahl, G. (2006). Group cognition: Computer support for building collaborative knowledge (pp. 451-473). Cambridge, MA: Mit Press.
• Watson, J. B. (1958). Behaviorism. Transaction Publishers.
• Wegerif, R (2011) Towards a dialogic theory of how children learn to think. Thinking Skills and Creativity. 6 (3) 179-195
• Wegerif, R., & Mercer, N. (1997). A dialogical framework for researching peer talk. Language and Education Library, 12, 49-64.
• Woolley, A. Chabris, C. Pentland, A., Hashmi, N., & Malone, T. (2010) Evidence for a Collective Intelligence Factor in the Performance of Human Groups. Science, September 30, 2010 DOI:10.1126/science.1193147

​Imagine if we were to get two robots – or chatterbots – each programmed to respond to words and categories of words with pre-prepared utterances and we made them interact, the results might look externally like a dialogue but it would not actually be one. Some so-called dialogues in social life can be a bit like that. It is quite possible for people to falsely claim ‘we are having a dialogue’ when they are just talking at each other or talking past each other. 

​Fortunately we know when social interaction is not real dialogue because we all know the experience of engaging in a real dialogue. Real dialogues happen when people listen to each other and learn from each other. Real dialogues tend to feel exciting and enjoyable. One way to characterise real dialogues, so as to distinguish them from mere external interaction of the robot kind, is to point out that in a real dialogue shared thinking occurs such that it is not always possible to say who is thinking. One could say that in a real dialogue there is no longer just ‘I am thinking’ and ‘you are thinking’ but there is also the experience that ‘we are thinking together’. 

The mutual resonance or entanglement or intertwining involved in dialogues results from a kind of circularity of reference that Rommetveit calls ‘attunement to the attunement of the other’ (Rommetveit, 1992). In communications theory, Rommetveit points out, messages go one-way from a sender to a receiver, whereas in a dialogue the process is more of a loop. The ‘other’ is always already on the inside of every utterance. This is because each utterance responds to what the other has said in a way that is intended to relate back to the other. 

Every theorist who can reasonably be referred to as dialogic addresses the dialogic space aspect of dialogue, but each does so in a different way. Bakhtin explicitly links dialogue to learning through his concept of the ‘internally persuasive discourse’ (1991 p376). Authoritative discourse, he writes, remains ‘outside’ us and remains static in meaning. You either have to accept it or reject it but you can’t really engage with it. In contrast the internally persuasive word or discourse is one that enters inside you as if it was one of our own words, it is ‘half ours half someone else’s’ and so it is able to re-organise our words from within and also to engender new words and new ideas.
 
In a similar way Buber contrasts the objectifying ‘I-it’ attitude that turns the other into an object with the intersubjectivity of the ‘i-thou’ attitude that engages responsively with the subjectivity of the other. The ‘I-thou’ attitude leads to entanglement which Buber characterises with a spatial metaphor, the space between or simply ‘the Between’ or ‘Zwischen’ (1958).

Buber extends the apparent inter-subjectivity of the I-thou relation to include relationships with non-human objects such as trees. This extension of dialogic relationships beyond human voices is also found in Bakhtin who remarked ‘I hear voices in everything’. It is essential to dialogic educational theory that we realise that the dialogic space is not only inhabited by the voices of physically present humans. In education the dialogic relationship from which we learn the most is often with a disembodied cultural voice. This might be a dialogue with someone like Shakespeare, for example, or with an area of discourse such as Mathematics. In each case Shakespeare or Mathematics could appear external and static, if the education is not dialogic, or, with a more dialogic education approach, they could dynamically enter inside a learner to become a living voice that helps the learner understand new things.
 
The idea of dialogic space was perhaps first introduced into the analysis of classroom talk when I was writing and thinking together with Neil Mercer about how to explain effective collaborative learning in small groups (Wegerif and Mercer, 1997). The issue at the time was how to understand social cognition in the way in which the upper primary children (aged 8 to 11) were talking together in small groups. There seemed to be at least three significant types of talk, disputational talk when children disagreed with each other without giving reasons, cumulative talk when they agreed without reasons and ‘exploratory talk’ where they genuinely engaged with each other’s ideas. We realised that each type of talk reflected an intersubjective orientation related to a form of individual identification.

​In disputational talk children identified with their own self-image or ego and each wanted to be the one to win the game and get the answer and in cumulative talk children identified with their image of the group as a harmonious unit and so did hot want to criticise. This seemed clear but then we asked ourselves, what is the form of identification involved with ‘exploratory talk’? A key feature of exploratory talk is being able to change one’s mind. The question then was, from what position is it that individual children are able to look at what they have said, find it wrong and so change their minds? This way of thinking about this practical issue led me to introduce the idea of identification with a ‘space of dialogue’ or ‘dialogic space’. 

While Buber’s ‘Between’ is always a highly abstract notion, 'dialogic space' has a concrete aspect. This was seen recently in a primary classroom in Japan where a group of three upper primary children were arguing about a puzzle presented on a tablet. Not only did their body language converge on this central focus but so did their fingers. Each put a finger on the tablet to point out what they thought the key to solving the puzzle was. Pretty soon it was clear that much of the shared thinking was being done by their fingers (research by Taro Fujita – in preparation to be published, similar to Wegerif et al referenced below).

Before dialogic space opens up things tend to be thought of as located in space using the default Identity Ontology that I wrote about in earlier blogs. Identity Ontology says that ‘a thing is what it is and not another thing’. The pepper and salt pots  on the café table are just pepper and salt pots. When dialogic space opens up then material things, bodies, hands, voices, gestures, pixels on the screen, become signs for other things and representative of voices that are not present. Depending on the dialogue the pepper pot could become Lionel Messy scoring a goal for Barcelona dribbling brilliantly around the salt pot and into a goal marked out by knifes and forks, or the two pots could represent the relationship between a proton and a neutron in an atom  (deuterium in this case) surrounded by an electron cloud of scattered pepper on the table or they could stand in for almost anything at all.

​Each opening of dialogic space is unique, but all dialogic spaces, once opened, share something in common. While in practice any given dialogic space might have a limited range of themes and probable outcomes these cannot be determined in advance because, in principle, any real dialogue opens up an infinite potential for creating new meaning. This is just another way of saying that the context that could be brought into play in any dialogue is not bounded, at least not as far as we know. In practice a group of children in a dialogue might just bring in references from the TV shows that they saw the night before but in principle anything could be brought to bear on the problem at hand. I often refer to dialogic space rather than dialogic spaces in order to draw attention to this unity of the structure of dialogic space as always opening up a potential for unbounded contextual meaning. But, of course, dialogic spaces are also all different in their physical location. Dialogues always open somewhere that could be given a location in space and time and also in terms of cultural and historical context. The neologism of ‘dialogic space(s)’ would be the most accurate term with the singular ‘space’ referring to the unbounded ideas side and the plural ‘spaces’ to the physical concrete side. 

Dialogues cannot be defined only in terms of the external surface of things: they cannot be reduced to visible countable things as if they were just so many words or bits of data. Wherever dialogue occurs it opens up a space of multiplicity and uncertainty - a dialogic space which consists of resonances between utterances and voices. The opening of this space is the precondition for the possibility of learning something new. To try to define dialogues in terms only of the external and the visible is to try to kill precisely what makes them not only useful but essential to education -  the internal and invisible dialogic space that makes new connections, new insights and new understandings possible. That is why we need a concept of dialogic space in education and in educational research.