In discussions of social life, the word ‘collaborating’ suggests people acting in a positive manner. Although a manner that is quite natural and uncomplicated. After all, the most routine of conversational exchanges often get approvingly termed “collaborative”. Such as deciding the best place to meet. Or discussing a possible gift for a friend. However, in terms of distinguishing different acts of learning, the word ‘collaborating’ is troublesome. The problem is that too many interestingly different learning practices can become blurred together by applying the same umbrella term. Take the following chapters of the present book for instance. They each refer to learning activities for which the description ‘collaborating’ would often – if not always – apply. So, a student and teacher in a tutoring session could be said to be collaborating. Even attending the exposition of a lecture can require collaborating (rules of social order get requested and respected). Nevertheless, it has to be admitted that such different learning practices do still have a common feature at their core. Perhaps the best way of expressing this is simply to say that they all involve people ‘thinking together’.
1. Introduction
The fact that thinking together features within many acts of learning will be repeatedly illustrated in this book. However, the present chapter highlights those learning circumstances where thinking together is the focal thing, where it is the very engine of what is being done. These are situations where we might say: “Let’s collaborate to understand this” or “Lets collaborate to get this built”. And they are situations where we suppose participation then promises learning. It is in the spirit of an ‘acts of learning’ approach, to discuss ‘collaborating’ by reference to its verb form, as something that we do. But understanding it as an act of learning will be greatly helped by including also its noun and adjective forms. These define two further windows onto the same topic. So, when we consider ‘the collaboration’ (noun), we will be considering alternatives for designing a form of learning episode involving multiple participants. When we consider ‘being collaborative’ (adjective), we will consider a particular state of mind, a particular disposition. (Rest assured that such slipperiness of word meaning is more pressing for this particular act of learning and not so urgent for those in later chapters.).
First, the noun: the collaboration as an educational design. If you and I discuss a possible gift for a friend, we would not always call that “a collaboration”. Doing so would too often elevate a casual conversation – albeit goal-oriented – into something more formal. Something more like the thinking-together episodes orchestrated in classrooms or workplaces: episodes that imply a kind of contract. That ‘contract’ entails students (or colleagues) signing up for something like this: partners are convened with matched expertise, a shared problem is defined, there is a commitment to democratic interaction, and a striving for closure – ideally with consensus and perhaps with an expectation of evaluation by outside actors (teachers or managers). It sounds a good way to solve problems. Moreover, joining in may be a good way to stimulate learning. Across this chapter, we term such problem-solving contracts ‘schooled collaborations’. Not that they are unique to schooling, but that is where they are usually first cultivated with determination. As it happens, psychological research on collaborating as an act of learning almost entirely addresses schooled collaborations. Such research has important questions to pose. How do children gain a confidence within such encounters? What benefits do they gain? And how does the design of collaborations support them in achieving benefit?
Second, the verb: collaborating. In relation to educational practice, this identifies the ‘democratic interaction’ within a (schooled) collaboration. A central theme of psychological research on this topic involves exposing, systematising and understanding what happens within such interactions. The success of a collaboration and the learning it might facilitate are best understood by exploring how partners talked and acted together. Then it becomes possible to understand how this collaborating evolves with experience and how learning outcomes are pursued with increasing confidence. Therefore, an important aim for research is to characterise those ‘thinking together’ strategies that best generate a productive outcome when collaborating.
Finally, the adjective: collaborative. If research does identify such strategies for productive interaction, then ‘collaborative’ describes individuals who regularly employ them. Pointing the research lens in that direction may inform a more wide-ranging sense of ‘collaborating’ as an act of learning. That is a sense of how experience gained in schooled collaborations can percolate up and guide the flow of any informal thinking with others. Considering the adjective form allows us to identify individuals or groups adopting a ‘collaborative approach’ to whatever everyday interactions have potential for learning. As well as influencing other formal study or instruction. That is, initiatives that are not always themselves formally termed ‘collaborations’, such as tutoring and exposition. In short, acting collaboratively is a disposition that individuals may bring to a range of learning situations – not just those problem-solving interactions that we formally label ‘collaborations’ (and concentrated on in this chapter).
The plan for our present route around these grammatical distinctions will be as follows. In Section 2 the schooled collaboration is considered as a designed encounter. Three examples are sketched in 2.1. While the evidence for the collaboration being an effective resource for learning is summarised in 2.2. Section 3 addresses collaborating as a social practice: how thinking together has been observed in schooled collaborations. The challenge of retrieving and processing knowledge with others is considered in 3.1. The role of conflict and convergence in conversation is considered in 3.2. That material is integrated in 3.3 where strategies for optimising collaborating are suggested. Section 4 considers the collaborative disposition or ‘being collaborative’ as a characteristic of individuals. Our deeply social nature is first illustrated through discussion of the human newborn in 4.1. The emergence of a theory of other people’s mental lives is reviewed in 4.2. Collaborative dispositions in the preschool period are considered in 4.3 as preparation for later schooled collaborations.
2. The collaboration
Building on the grammatical distinctions above, this section discusses collaboration as noun; as an arrangement for learning. That means recognising a certain kind of design for thinking together – thinking towards solving a problem, creating an artefact, building a perspective, reaching a decision or, more generally, striving towards some shared goal. The first sub-section (2.1) illustrates typical designs for bringing students together in this spirit, typical configurations for a learning opportunity. The three examples sketched also allow illustration of the research methods for understanding the outcomes of such encounters. The second section (2.2) reviews their outcomes in terms of learning.
2.1 The collaboration: Three examples of its configuration
The noun ‘collaboration’ points to a circumscribed event. The present chapter considers the learning that is possible within such ‘schooled collaborations’. Of course, collaborating happens elsewhere than schools. But our concern is with learning and so groups configured in educational (or workplace) contexts are of special interest. Besides, classroom experience doubtless grounds our confidence and motivation for thinking with others in many other situations and at other times later in life. To anchor the discussions in this chapter, three examples of schooled collaborations are sketched below. They come from published reports which, taken together, illustrate typical research methods. But also, they illustrate the range of topics commonplace in schooled collaborations. In these examples, the topics are physics, historical interpretation, and building a physical object.
The first case (A) clearly addresses a central concern: namely, how to judge whether a collaboration has been effective for learning. This requires comparing the outcome for individuals who studied by themselves (‘solo’) with the outcome of (different but matched) individuals who each studied as one member of a collaboration. Chi and Menekse (2015) describe an example for university physics:
The two conditions were 10 pairs of undergraduate partners solving kinematics problems with a text and 10 solo undergraduates solving the same problems with the same text. The participants had to learn the first four chapters of the text and the conceptual part of the fifth chapter to a criterion. Then they were given a pretest that assessed their understanding of nine concepts related to kinematics. After the pretest, they were randomly assigned either to solve three problems taken from the fifth chapter with a partner or to solve the three problems alone. Finally, they took a post-test that was identical to the pretest. (p. 258-259)
This is just one study. What happens across the range of such studies is considered in the next section (as far as this example goes – the learning gains were greater following collaboration). Note how the method works. Each participant’s baseline knowledge is established through a ‘pretest’. Then, after studying, each student’s level of understanding is tested again – to evaluate learning. One point of method: the pretest must be challenging, because if students manage it too easily any improvement after learning may not show up when the test is taken again.
In that widely-employed research design, learning is sought as a state change: an alteration in what participants know, as assessed by before-and-after testing. However, participating in a collaboration can also be treated as a kind of performance: simply documenting and characterising what students actually do when they are convened into thinking together. After all, a schooled collaboration is a way of interacting that has to be grown into. It differs from solving problems together in the playground. In the classroom, the problems tackled are typically imposed, partners may not be those preferred, the place and duration of the interaction are constrained, a democratic attitude to participation is expected, and there may be a responsibility to report what was found and then have it evaluated. In short, thinking together in class is shaped into a particular kind of cultural practice. We can research how willingly participants engage with such arrangements and what strategies they adopt when doing so. These matters are easily studied, because collaborators must make their thinking shared and public – including to researchers who can look on. In this second example, (B), Pontecorvo and Giradet (1993) were looking on while 30 9-year-old students conducted a collaborative argument. The collaboration goal was to resolve typical uncertainties that will arise when interpreting an historical narrative.
Children were asked to discuss together and reach a shared judgment about the interpretative description that Ammiano Marcellino (a Roman historian of the 4th century) gave of the Huns … The task was preceded by curriculum activities (guided by the teacher and lasting for about 10 hr of lesson time) involving the critical reading of historical documents concerning German populations in their relations with the Roman Empire … The children were asked to reach a consensus on: (a) what the historian meant by “habits similar to beasts…” (368-369) [Further such tasks are listed after this one]
This study made visible how students achieved an interpretative argument about historical events. The researcher’s analysis demonstrated that the children’s autonomous discourse was: “…often on a higher cognitive level than that guided directly by the teacher” (392). So, a report of this kind offers an awareness-raising case study that can influence practice. It is a kind of ‘natural history’ study of the collaboration. Moreover, the observed interactions can be systematised into categories of conversational move. Such an analysis resources further studies that can evaluate which of those moves correlate best with effective learning outcomes. Similar research will be highlighted in section 2.2 below, where it contributes to understanding how collaborations ‘work’.
In our final example (C), the goal of collaborating was to build a LEGO structure. As with case ‘B’, the researchers were not chasing learning outcomes with a pretest/postest design. Their interest was in the membership composition of a collaborating group and how its mix might differentially shape communication. Jiang, Zheng and Han (2017) focus in particular on composition by gender mix. They describe their method as follows:
In this experiment, the 12 [4th grade] participants engaged in LEGO activities after school four times in total, once per week. For each activity, the teacher created a situation and then proposed a problem that the participants needed to solve collaboratively. Each group discussed and constructed an artifact together using LEGO bricks to solve the problem in about 40 min without any intervention from the teacher. The process by which the students collaborated to solve the problem was recorded on video. Then, the groups demonstrated how their artifacts solved the problem in the created situation (131-132).
Example C again illustrates an interest in capturing the performance of collaborating. In example ‘B’, that interest was cognitive: asking, how did a process of reasoning take place given the social composition of the group. In case C, the study addressed the significance of different gender mixes.
Our concern is with collaborating as an act of learning. However, two of these three examples involve no assessment of learning – at least not by a familiar before-and-after test procedure. Moreover, for the case of building with LEGO (‘C’), it is unclear what a measurable learning outcome might be – at least of curriculum-related knowledge. What such studies actually ‘measure’ (or, rather, ‘systematise’) is how students are observed to behave when asked to think together: deploying talk and action to get towards some goal (a LEGO object, an historical interpretation, a concept in physics, etc.). Communication strategies observed this way can then be usefully correlated (perhaps in later studies) with learning outcomes and thereby explain how those outcomes are best achieved. But these observations concern ways of acting that themselves change over time. Because a student’s style of thinking with peers is something that will evolve through increasing exposure to schooled collaborations. In short, how collaborating is ‘performed’ is something that matures. It is through documenting that process that the design of cases ‘B’ and ‘C’ do concern matters of learning. Not learning made visible by before-and-after assessments applied to a single episode of collaborating. But learning that becomes visible when it is time-tracked (or age-tracked) to reveal developmental changes in strategies for thinking together. Much of this chapter is about learning from collaborating, but learning traced in the manner of cases ‘B’ and ‘C’ is learning about collaborating. What is being learned is something referred to in the opening as a ‘collaborative disposition’ – the adjectival form of collaborating- and it will be re-visited in Section 4
Each example above illustrates a collaboration as an act of learning. Its effectiveness depends upon furnishing a ‘problem’ that demands solution (jointly) by exercising or extending some realm of knowledge. Moreover, the range of targets for such problem-solving is wide: things are constructed (LEGO), interpretations are made (historical events), and symbols are manipulated towards a solution (physics). But we must ask: how effective are experiences of this kind as acts of learning? The efficacy question for experiences like these is addressed next.
2.2 The collaboration: Evaluating learning outcomes
The experience of learning with others can be arranged in many different ways and the term ‘collaboration’ is applied quite loosely across them. So, judgements regarding the efficacy of collaborative learning depend on how the researcher has framed their concern. Much attention has been given to the example of whole classrooms (or whole schools) that are pursuing a general policy of group working. Settings termed ‘cooperative learning’ are the most closely researched (Johnson and Johnson, 1987; Slavin, 1996). However, ‘cooperative’ can mean a very broad commitment to group projects, without defining specific strategies for managing them. In general terms, it refers to: “…the instructional use of small groups in which students work together to maximise their own and each other’s learning” (Johnson and Johnson, 1999, p.73). If the efficacy of collaborating is judged in these whole-class or ‘policy’ terms, then reviews and meta-analyses report that learning outcomes are successful for such settings. That is, outcomes are more favourable than those from comparison classes/schools that centre instruction on cultivating individual study strategies (Gillies, 2014; Kyndt, Raes et al, 2013; Pai, Sears and Maeda, 2015, Roseth, Johnson and Johnson, 2008). That said, whole class working can be managed in a range of ways and success can be variable across different implementations (cf. Stanczak, Darnon et al, 2022).
Research in such settings provides what might be termed ‘aerial views’ of thinking together. However, an act of learning approach is concerned with arrangements at a finer level of granularity: research addressing the outcomes from individual episodes of collaborating (schooled collaborations). Intuitively, we probably feel that such collaborations must make better contexts for a person to learn in. That expectation is enshrined in the slogan: ‘two heads are better than one’. And perhaps for some of us, the experience of working in classrooms has reinforced that everyday belief. For sure, many practitioners will declare that collaborating groups generally do produce better work than individuals working alone. Moreover, research studies confirm this. However, what is confirmed is a somewhat limited claim about collaborating. Namely, a belief that the singular output of a students’ collaborating exceeds the typical level of work produced when those students work solo. Perhaps that is the comparison implied by our casual endorsements of collaborating. However, it is not a very telling comparison. For instance, groups may have a most competent member – an individual who intervenes and directs things to ensure a strong joint product. In such cases the group as a whole may indeed do very well (better than the average of solo students) but it may mean that some group members learn less from a procedure they were not fully part of. Similarly, a group may creatively share out responsibility for their task, taking advantage of the distribution of participants’ expertise. Again, this is clever strategy but it may undermine the extent of learning by some participants, according to their role in the total effort. To better judge collaborating as an act of learning, we require a different research method.
Ideally, this would involve the following. Convene a large group of learners and assess the knowledge of each participant on the topic about to be studied. Then separate them into those who will study in a collaborating group and those who will study solo. After the study period, evaluate each participant’s knowledge level again. This allows a judgement as to whether the average achievement of collaborating individuals is greater or less than the average achievement of solo-learning individuals (cf. Case ‘A’ in the section above). That method provides a strong basis for venturing a generalisation. However, for maximum confidence: repeat such comparisons over and over! Repeat the procedure for different study topics, and for groups of different size, age, composition, knowledge levels, motivation, etc.
Many studies repeating the above core procedure have been published. Casually browsing this literature will convey the impression of findings demonstrating that an individual student learned more after studying in a collaboration – compared with students who had studied solo. However, interpreting results from multiple published studies faces a recurring problem: namely, a possible publication reluctance (or submission reluctance) for studies that sought effects from that comparison but failed to find them. Moreover, apart from such unpublished, no-difference findings, a browsing of research must also exclude unpublished studies where the collaboration participants simply would not engage (for an example of that sort that was published, see Baker, Bernard and Dumez-Féroc, (2012)). A substantial commentary by Baron (2003) clearly identifies cases failing to confirm our intuition of an advantage for learning in collaborations. Certainly, Baron’s review gives no grounds for supposing that a collaboration will always serve its members better – that is, when compared to studying alone. Therefore, it may not be a matter of whether collaboration better supports learning but when and how it does (Nokes-Malach, Richey, and Gadgil, 2015).
Yet that should not inhibit a search for trends in the findings evaluating collaborative learning. The most robust conclusions must come from reviews that have scrutinised as many such studies as possible. A meta-analysis of that kind by Tennebaum, Winstone et al (2020) is particularly instructive. They concluded that the direction of research findings for peer collaboration was strongly towards students experiencing positive learning outcomes – when compared with other types of learning arrangement. This applied across different gender and age groups and there was no influence from the type of task set. Although collaborating was not shown to be more effective when compared with tutoring (an act of learning to be discussed in Chapter 2 of the present book).
Finding favourable trends and directions do not conclude a story. If a collaboration works, how does it work? A collaboration remains something to be understood and managed as a process of communication: something to be shaped towards being the most productive and engaging learning experience possible. To those ends, we need to understand the psychological processes that underpin effective learning with others. These will be explored in the next section (2.2) where consideration is given to collaborating (verb form) as a particular way of interacting with others.
3: Collaborating
The best prospects for understanding how collaborating ‘works’ will depend upon looking more closely inside collaborating exchanges – looking at the process of collaborating as a way of interacting with others. In 3.1 below collaborating is considered as a way in which our thinking mind can be reinforced (or ‘extended’) by a reciprocity of thinking with other people. We find that collaborating can both facilitate and inhibit thinking (3.1.1) and how those effects relate to thinking the generative thinking that underpins effective learning (3.1.2). In 3.2 the structure of supporting generative thinking during collaboration is grounded upon a commitment to social coordination but also is animated by episodes of conflict and its resolution.
3.1 Collaborating: A (socially) extended mind
The phrase ‘extended mind’ may not be a familiar one, yet it surely offers a good fit to the experience of being in a collaboration. It will be adopted here to explore what happens when people set about solving a problem by thinking together. The phrase offers a useful starting point, not just because it feels right but because it also happens to have some standing in cognitive psychology, having recently evolved there as an innovative way of theorising mental life (Clark and Chalmers, 2010; Rowlands, 2010). For a long time, Psychology has addressed the mind largely through a common-sense model: a traffic of operations carried out ‘in the head’. There was always a risk that such a perspective on cognition implied a Sherlock Holmes version of intelligence (chair, pipe, pondering). After all – if we are honest – we do not relish those self-contained episodes and they can be hard to sustain. Besides, while Holmes’ manner could be muted, his intelligence was actually much more versatile, much more in-the-world. It might even be judged an optimal version of what the rest of us attempt. He was keenly attentive to events around him – which he interrogates – he had a powerful memory for past experiences, and a ruthless capacity for logical inference.
However, there is one aspect of modern thinking practices that seems absent in Holmes’ intelligence. He rarely annotated his actions (but see The Adventure of the Devil’s Foot for an exception). Such practices are surely an important part of what intelligence entails. Because external records provide a form of memory around which reasoning and prediction can be organised (Allen, 2024). They could be termed ‘tools to think with’. That phrase nicely captures what psychology means by referring to the mind as ‘extended’. It is a perspective that allows ‘thinking’ to mean more than only private computations ‘in the head’. The extended mind entails computations that are instrumental in the world: they are executed through a collaborative relationship with the environment. What we traditionally regard as the mind – that internal space of thinking – becomes ‘extended’ whenever we integrate its computations with resources that are external to the self. Doing so is sometimes termed an ‘offloading’ of cognition (Risko and Gilbert, 2016). So, for example, if the product of thought requires juggling numbers, we might offload to a calculator. If it requires finding a route, we might program the GPS. If something must be remembered, we might write it down. At these everyday moments, the computations of thought (the calculating, the navigating, the remembering) are pursued through our seamless engagement with environmental tools.
Other people are often part of our thinking environment. At the risk of sounding ruthless, those people (when collaborators) could also be recruited as ‘tools-to-think-with’. Sherlock Holmes’ rather cerebral intelligence may again seem an imperfect example but, nevertheless, Holmes’ companion Watson does serve him in that way. It might do so when together they try to recall significant events, or when some line of reasoning can be tested by jointly exploring it. (In The Hound of the Baskervilles Holmes remarks to Watson “it may be that you are not yourself luminous, but you are a conductor of light”.) The extended mind model has been useful for better understanding our in-the-world intelligence. So, it can be applied to collaborations: situations where the mental ‘extension’ involves actively integrating our minds with those of other people. Within which contexts the phrase ‘socially extended mind’ may be applied. Adopting it directs attention to the interlocking of two or more mentalities; it frames collaborating as a dynamic system, a conception more open to research than the once-fashionable ‘group mind’ (see Pavitt 2003 for a critical treatment of that).
One way of expressing the prospects for such a dynamic system of shared thinking is through the mantra that often underpins our faith in collaborating: ‘two heads are better than one’. It’s a simple claim and one that does hold up across all sorts of problem-solving encounters (and, informally, is termed ‘the wisdom of crowds’). If its message seems obvious, it is because the mantra makes a rather limited contrast. It merely asserts that two people thinking will generally achieve more than one person thinking. Yes, more brain power is available! However, a more useful comparison comes from asking whether the consequences of two heads thinking together are always better than the consequences of two heads thinking alone. Or, representing this as a practical challenge: what is the best way of configuring a small group of students to be most productive – group them to think together or separate them to think alone? If ‘productivity’ refers to the learning achieved by the participating student ‘heads’, then the conclusion of Section 2.2 is already on the table. For learning, it is usually better to organise the heads into working together: i.e., to arrange a collaboration. Yet learning is not always the primary purpose for structuring a thinking exercise and so learning is not the only outcome that invites research. Another purpose is an arrangement sometimes termed ‘brainstorming’. There the issue becomes how to manage individuals so that their combined thinking creates the best (or most) collection of certain ideas that the organiser has requested. For example, producing designs for some specified gadget, or a set of possibilities for a colleague’s retirement present, or making a list of American state capitals, and so forth.
Researching productivity for that kind of purpose is not a matter of asking whether individual brainstormers learn anything (although they may), still less is it about the best way for them to do so. Studies will concern how and when the quality and/or quantity of ideas from heads thinking together is superior to the aggregate of ideas from heads who thought alone (an arrangement which researchers term a ‘nominal group’)? What research generally reveals is that the productivity of ideas from brainstorming heads working together is inferior to the productivity of the same number of brainstorming heads working alone. (In these comparisons any repeated material from the heads-alone (i.e., the nominal group) is treated as redundant and counted only once.) This comparison therefore implies that brainstorming groups who are thinking together must be failing to capitalize on the potential capabilities of their members. Of course, that is an observation more likely to be of interest to workplace managers, rather than educational practitioners. Because, in the workplace, group productivity of ideas is likely to trump individual learning as a priority. But the findings from brainstorming research will turn out to be relevant to both. So, we must look more closely at what happens within the socially-extended mind both when it brainstorms and when that brainstorming is directed towards learning.
Joining in with others cannot always be the best bet for effective learning – there can be numerous practical constraints of time, place and purpose that will influence it as a choice over learning solo. Yet, other things being equal, collaborating still remains a well-proven way for individuals to learn (2.2). The apparent underperformance of brainstorming groups is not such as to challenge that claim. Yet it is intriguing, because even if what we term ‘brainstorming’ is not the same as what we term ‘collaborative learning’, the activity of the former is surely often some part of the latter. For example, a collaborative learning group will often have to remember together facts relevant to their set task. Part of what they may therefore do is what we would term brainstorming their collective memory. In which case, studies have shown (e.g., Rajaram and Pereira-Pasarin, 2010) that remembering conducted collaboratively generates less output than the sum of output gathered from the same number of participants remembering solo (again, excluding redundantly repeated items). Similarly, another domain researched for brainstorming – and one relevant to collaborative learning – is the generation of creative ideas. The same discrepancy is found there: ideas collected up from a nominal group (excluding repetitions) are typically superior (in quality and quantity) to ideas generated by a matched group working together.
These results are commonly said to reveal what must be an otherwise hidden process of ‘collaborative inhibition’ (Barber, Harris et al, 2015; Basden, Basden et al, 1997). It may help to express the meaning of this inhibition in more concrete terms. If 10 individuals thinking alone generate 100 ideas then it is surprising that a group of the same (or matched) 10 individuals thinking together come up with less than 100 ideas. Such inhibitory effects of collective thinking are evidently not so powerful as to undermine the overall facilitatory advantage of collaborating as an act of learning (see 2.2). Nevertheless, something unexpected is going on. Perhaps something whose inhibitory effects on collective thinking may reduce the scale of its facilitatory advantage for learning: and so, learning collaboratively may not always make as big a difference as was hoped for. Perhaps not as deep as it could be, or perhaps not as inclusive of its participants as it could be.
3.1.1 Two ways collaborating can inhibit the extending of mind
That collaborating can have inhibitory effects on the experience of those taking part is an inference drawn from studies of groups. In such research, the productivity of a group brainstorming together is compared (unfavourably) with the productivity of a group of solo brainstormers. Further research is necessary to determine the nature of the inhibitory forces that seem to be active in collaborative thinking. These will be identified in the present section. While in the next section (3.1.2) it will be explained how their presence can constrain the level of learning that we expect from collaboration – either in terms of the depth of what the average learner achieves, or in terms of the spread of achievement across participants in the collaborating group. It is important to identify the nature of these inhibitory factors, because understanding them may help improve how learning groups are best convened, prepared or managed.
Research now indicates that collaborating can exert inhibitory effects on thinking through at least two routes. One might be termed ‘social’: it concerns the limited engagement of some participants when a task is made collective. The second is more ‘cognitive’: it concerns disruptions to the thinking strategies adopted when participants find themselves processing information simultaneously.
First, limited participation. It has been a recurrent finding that individuals working in a group invest less effort in striving for a goal than they would if working alone. The grouped nature of some activity can inhibit applied effort. This was first made empirically visible over 100 years ago in a study by Maximilien Ringelmann. He monitored the effort levels of individuals while they were pulling together on a rope. The measured effort of an individual when part of a group (i.e., in a tug-of-war) was less than the effort exerted by that individual when acting alone. Moreover, the larger the group, the greater the reduction of effort. This ‘Ringelmann Effect’ was later replicated by Ingham, Levinger et al (1974) – and others thereafter – for a variety of tasks. Moreover, in any given group task, this lowered effort may be more firmly adopted by some individuals than others. Psychologists have termed such managed disengagement ‘social loafing’ (Gabelica, DeMaeyer and Schippers, 2022). In effect it implies that designing a task to be collaborative can have inhibitory effects on the participation of some, if not all, of its members. The requirement to think with others allows the uninspired participant to disengage from a task designed for learning. For some, task engagement can be inhibited by collaborating; their effort is offloaded onto partners.
However, research has identified other ways that inhibition can be expressed as disengagement from tasks – ways more subtle than deliberate loafing. Partners will bring different personalities, different motivation, and different vulnerabilities to a collaboration (e.g., Baron, 2003; Bell, Brow et al, 2018). In such cases, collaborating can inhibit task participation by arousing self-identification concerns. For instance, individuals may withdraw because they decide that their input is not sufficiently valued by others. Or because they lack confidence in the quality of the contributions they feel able to make. Or they may merely contribute material already known by partners in order to receive approval or reassurance from the group. Such fragile engagement can mean that those individuals may learn less from their participation. If such disengagement is not anticipated and repaired, the potential of collaborative learning may be underestimated.
The second factor behind collaborative inhibition concerns how the mind – when socially-extended – collectively processes information: how information retrieval, sharing, and application are managed when different minds are required to do these things together. Consider collaboration case ‘B’ in Section 2.1 above. Students were required to construct an historical interpretation. This must surely involve some collective remembering: historical events and actors need to be recalled, judged for relevance, and entered into the thinking mix. Some researchers (e.g., Jeong and Chi, 2007) suggest a Venn diagram can describe what is involved. Each collaborator has their own ‘circle’ of knowledge and, to some extent, those circles overlap. However, the overlap has to be discovered (and expanded) through social interaction, in order for it to become the ‘common knowledge’ that is then worked upon. Yet that ‘discovering of overlap’ – when rendered a shared commitment – may be imperfect.
To be more specific, consider again the history example. It might concern a topic such as: ‘King Henry VIII’s role in establishing the Anglican church’. Group partners may riff their remembered ideas off each other, taking a path such as: wives > Bolyen > divorce > Wolsey > Pope > authority > Reformation > monasteries…etc. Collaborators loop forward in a shared tip-of-the-tongue manner, recruiting each other’s prompts to remember things – often beyond what they could each recall alone. Such a routine is useful: it delivers every participant a strong factual grounding for the interpretation that the set task requires. It illustrates the common-sense expectation that listening to someone else recalling material can usefully facilitate one’s own recall. However, it can also impede or deflect that effort. Informally, this might be expressed by someone saying “your suggestions made me lose my train of thought”. Put more psychologically, this has been termed ‘retrieval strategy disruption’. it is not a necessary consequence of interactive remembering. But a meta-analytic review of relevant research confirms that it can happen and it can form one basis for a certain underperforming in a collaborating group’s recall of information (Marion and Thorley, 2016).
Consider this disruption of memory more deeply, again using the history task as illustration. The remembered ideas have to be integrated into an historical interpretation. But now the productive integration of data can also be inhibited by the collective nature of the thinking. Certainly, other peoples shared creative ideas can prompt the imagination of their partners, but working collaboratively can also be disruptive of that process. For example, a study by Kohn and Smith (2011) described collective thinking in which undergraduates generated ideas either in collaborative groups or in matched nominal groups (i.e., thinking alone). As found elsewhere, the nominal groups outperformed the collaborating groups in quantity and quality of ideas. Close observation of these collaborating sessions made sense of the shortfall. Collaborators were more vulnerable to ‘idea fixation’: that is, an over-conforming to, and persistence with, those ideas that arose early in discussion. Such fixation has been shown to be less likely when thinking alone, and less likely if a group is allowed (or encouraged) to switch between different and independent thinking tasks (Lu, Akinola and Mason, 2017).
To summarise Section 3.1: being in a collaboration generally facilitates thinking but it can also have inhibitory effects. One form of inhibition arises when the effort of collaborators becomes an excuse for making limited effort oneself. A second form arises from the direction of thought taken by others: it may often inspire, but it can sometimes distract, replace, re-route or prematurely fixate one’s own thinking.
Yet perhaps such inhibitory processes are benign in relation to learning? In the end (and as stressed above), the learning advantage for collaborating over solo study still holds up well for the average learner. This is so, despite the potential of collaborating to inhibit some members from taking part. Or despite the occasional inhibition of recall and reasoning arising from simultaneous information processing. Perhaps such factors can seem benign in managing group work because we often say of collaborative learners: “all’s well – they did get there in the end”. Yet collaborative inhibition may start to appear more troublesome when considering the ‘getting there’ experience of individual participating learners. Has the learning occasion that is collaborative fully equipped them with an analytic method for the problem – as well as a conclusion? This will be considered in the next section. But doing so requires departing from the terms of the discussion so far: shifting from exclusive attention to the socially-extended mind (collective thinking) towards incorporating the learner’s internally-contained mind (private thinking).
3.1.2 How collaborative facilitation and inhibition shapes generative thinking
In Chapter 1 (Section 3.1) it was noted that psychological theory has become increasingly attentive to individual agency as a factor in learning. More particularly, theory supposes that learners need a generative relationship with the substance of their learning episodes. The following quotation was used earlier to illustrate this concern: “Generative learning involves actively making sense of to-be-learned information by mentally reorganizing and integrating it with one’s prior knowledge” (Fiorella and Mayer, 2016, 717). This seems a straightforward idea. However, taking it forward in relation to collaborating requires probing further what is meant by the “mentally reorganising” of incoming information or “integrating it with one’s prior knowledge” – and, more particularly, what happens when asked to do those things with other people. Such concerns require understanding how cognitive psychology conceptualises “one’s prior knowledge” as a mental structure.
Any sketch here of our mental apparatus – and an exploration of how it functions within collective thinking – will benefit from a working example. So, consider for instance the collaboration topic in history education that was invoked above: King Henry VIII and the Anglican church. It will be enough to highlight the fate of just one element in the collective thinking narrative imagined earlier: let’s take the element ‘Wolsey’ (an advisor to the King).
Psychologists propose that an individual’s prior knowledge is stored in a stable structure termed ‘long-term memory’ (LTM). The architecture of that stored knowledge is theorised to be a network: an individual item such as ‘Wolsey’ may exist as a node in this network. It will link with other nodes (Anne Boleyn, the Pope, etc.) with varying strengths. (Note that the networked configuration of nerve cells in the brain have helped make this model compelling.) The patterning of networked links serves to code what we would casually term a concept’s meaning (a concept such as ‘papal authority’). Accordingly, the meaning attached to any particular concept will vary between individuals. Meaning will vary because individuals have different histories of interacting with information sources and thus different histories of integrating new information with their existing networked patterns in LTM. This ‘integration’ involves a new input activating an appropriate node in this network such that, by a process of ‘spreading activation’, related nodes will also be activated and their relationships incidentally strengthened.
Thus, when it is proposed that integrating to-be-learned information can be a ‘generative’ process, it is being proposed that the learner takes more conscious responsibility for managing that spread of activation. And the input of collaborators can prompt that. For example, suppose after ‘Wolsey’ has been introduced in the conversation, some collaborator comments: “he was one adviser that wasn’t executed!” An attentive (and generatively thinking) collaborating partner might then activate related concepts in the network ‘neighbourhood’ for their concept ‘Wolsey’ (perhaps activating links with Cromwell and More, who were executed).
That generative practice of actively exploring and integrating one’s prior knowledge is a process executed in ‘working memory’ (WM): which is the second important mental structure proposed within psychology. Working memory is the mind’s processing engine. It manages elements retrieved from LTM, perhaps integrating them with input of the moment – an input such as a collaborator saying “Wolsey was executed.” WM is fragile, with limited processing space and with contents that have limited duration if not rehearsed. Therefore, people collaborating can usefully furnish a single, shared working memory – overcoming the limitations of individual WMs. So, a collaborating partner suggesting “Wolsey?” can prompt someone else to make (and share) a link such as “Papal Legate!”, and so on extending the historical narrative. The socially distributed nature of this collective information processing can mean that the individual learner is relieved of some WM processing effort that would have been necessary if they had been working alone.
As sketched above, the architecture of LTM and WM offers a compelling framework for hosting the depth of processing that a generative response to new information entails. Although in more familiar language, ‘generativity’ could be expressed as what we naturally term a ‘thinking through’ prompted by some idea cropping up during a discussion (such as “Wolsey!” when discussing Henry VIII). This in turn may resonate with our everyday intuition that the ‘journey’ of thinking can be as important as its destination: because on such journeys our already-stored knowledge gets exercised, linked, and elaborated. That is to say, we enact a generative process of thinking. Against which background, an important question for collaboratively-organised learning then becomes: How does thinking that is made socially-generative distinctively shape the learning of those individuals participating? Is it universally helpful to pursue generative thinking with others? Or, expressed on the metaphor of travel, how might a collective thinking journey variously facilitate and inhibit our own personal progress towards a learning ‘destination’?
It has been explained above how collaborating creates socially-distributed and outwardly-manifest versions of LTM and WM. As such, they will be shared spaces for both generative remembering and generative reasoning. The advantage of being able to share resources in this way must help explain the successful outcomes from learning that is collaborative – as summarised in 2.2 above. But in that summary, it was also acknowledged that, when compared with solo learning, collaborative learning was not always a success. Moreover, even where it is shown to be successful, its advantage sometimes appears more modest than common sense anticipates. If the outcomes of collaborative learning are sometimes disappointing, this may arise from how the collectivity of collaborating can restrict the individual participant’s control over a generative thinking process. For example, the social loafer may opt out of doing such thinking when the effort is being taken by others. They may therefore take away the conclusion of the discussion (and so appear to have learned perhaps) but not have been active in the process of reaching it. Similarly, participants dis-engaged through their own social insecurity may be denied generative thinking because they concede responsibility for it to others in order to protect their self-confidence. While, on the other hand, an over-confident participant may constrain the generativity of their partners’ thinking by cultivating conforming or ‘herding’ towards that participant’s contributions (Morgan, Laland and Harris, 2015). Moreover, even when task engagement is healthy and expertise is balanced, the groups generative thinking may be vulnerable to a fixation of attention to those ideas that occur earlier in the discussion.
A credible prediction follows from a collaboration potentially fracturing the experience of thinking generatively: namely, that sharing the journey to a problem’s conclusion without generatively deriving it may foster overconfidence in what has been learned. As it happens, collaborators have been shown to have more faith in what they have learned from thinking together than is subsequently found to be the case (Minson and Mueller, 2012; Koriat, 2012). This in turn may lie behind findings such as those of Sampson and Clark (2009). These researchers studied high school students working on science problems. After their group work, one-third of these students produced individual explanations that were inferior to the solution submitted by their own small group.
However, there are strategies for anticipating the learning overconfidence that some individuals may take away from collaborating. For optimal learning (and subsequent confidence), participants need to be fully involved in the (socially) generative thinking required by their assigned task. When Chi and Menekse (2015) correlated successful learning outcomes to different patterns of collaborative dialogue, they found that the greatest success followed discussion where collaborators were mutually constructive. The researchers defined this form of dialogue in terms of participants exercising together the generative approach to remembering and reasoning that has been highlighted here. It is in the gift of participants to animate their dialogue in that way. But, equally, in any collectivity of thinking participants will benefit is there is protection from the various forces of disengagement, dominance and fixation that have been outlined above as fracturing generativity of thinking. The collaborating group needs to maximise the cognitive participation of all its members. To some extent this can be achieved by teachers and managers convening collaborations with ‘‘contracts’ in which such expectations are somehow specified and legislated at the outset (Kameda, Toyokama and Tindale, 2022). But protection will also depend upon a participatory dynamic that is sensitively managed by group members themselves. The observations on generative remembering and reasoning made above have stressed the importance of collaborating partners owning the thinking taking place. A successful group is likely to be one where both the cognitive and the social dynamic of participating have been sensitively manged by its members. Such interpersonal responsibilities that arise during collaborations are a topic that heads up the following chapter section.
3.2 Collaborating: coordination and conflict
Exchanges that construct the socially-extended mind (3.1) are sometimes described as ‘transactive’. The term expresses collaborating partners striving to interlock their thinking. It has been defined as: “The shared division of cognitive labor with respect to the encoding, storage, retrieval, and communication of information from different domains that often develops in close relationships” (Hollingshead, 2001, p. 1080). Such shared division of thinking is surely familiar from our everyday communication within relationships that are close for us. But such productive harmony can be difficult to achieve in classroom collaborations. And even more so in research exercises. Some researchers have themselves commented: “…when very short experiments are conducted, involving students who neither know each other, nor are given the opportunity to know each other, the question of a developing interdependence is a mute one” (Salomon and Globerson, 1989 p.93).
There are numerous studies illustrating the challenge of this social coordination during collaborative learning. A collaboration inevitably brings together individuals who have their own goals, personalities and emotions. Progress towards them learning together will depend on how well they create that ‘closeness’ of relationship. Naykki, Jarvela et al (2014) provide an example of the challenge. Observing undergraduates collaborating over a 3-month period, their study illustrates how unresolved motivational and socio-emotional challenges within groups can undermine learning. Lacking effective coordination, students were more likely to withdraw from interaction or lower their goals – rather than working to defuse discord. Such threats to progress require effective personal and social regulation mechanisms (Jarvela, Malmberg et al, 2021). Indeed, sometimes the management of relational matters may dominate the collaborating. For example, Janssen, Erkens et al (2012) studied 101 groups of secondary school students working together on history tasks in a networked environment. The regulation of the group’s communication represented a major proportion of their effort – yet doing so did correlate with effective group performance. In particular, these authors identified a significant concern among participants to manage what can be termed the ‘grounding’ of communication. That is, interactions asserting, questioning or confirming the groups’ understanding of their task.
Many researchers suggest such convergence of thinking is key to the definition of collaborating. For example: “…coordinated, synchronous activity that is the result of a continued attempt to construct and maintain a shared conception of a problem” (Roschelle and Teasley, 1995, p. 70). However, the ability to achieve such synchrony of purpose – to converge on a problem solution – is greatly helped by exercising a distinctively human capability: namely, the capacity for intersubjectivity. This is a concept central to collaborating but which also recurs in the chapters that follow, so it needs to be defined here.
Subjectivity refers to an individual’s momentary cognitive or emotional state. Inter-subjectivity is then an achievement of mutual awareness between people regarding their respective subjectivities. It underpins a capacity sometimes termed ‘mindreading’, because its consequence is a drawing towards ‘reading’ the subjective states of others. This is a capability that is deeply human. Indeed, the ability to strategically coordinate with others in this way – in effect, to collaborate – is commonly invoked when explaining human evolutionary success. Consider the management of a simple hunter-gatherer activity such as an ambush. Each individual must be sensitive to the perceptions, actions and intentions of their partners (while electing to hunt in such partnerships ensures protection in numbers). In short, intersubjectivity will underpin such effort as the effective ambush by making possible goal-coordinated action (i.e., collaborating). In the present context it is identified as central to our ability to learn with others in contexts of education.
The harmony associated with social coordination is an overarching interpersonal force governing effective collaboration. But there is an opposing force of conflict that is also significant in how we learn through thinking together. Jean Piaget is an influential theorist who stressed the constructive role of conflict. Collaborating became regarded by many as a valuable act of learning because of its capacity to promote social conflicts around differences of understanding. Such conflicts invite resolution and, so theory insists, the effort of resolving conflict resources learning. However, the conditions for such effects can be fragile. A disagreement does not promise a productive resolution. In the face of disagreements, collaborators can be skilled at gently disposing of them rather than exploring them. For example, several research projects have illustrated how they can evoke reactions such as rejecting, ignoring, excluding or temporarily parking an issue (Barziliai and Ka’adan, 2017; Chinn and Brewer, 1993; Kienhues, Stadtler and Bromme, 2011).
Yet it is easy to understand how collaborative disagreement can have a productive influence, rather than a corrosive one. On this expectation, useful and shared understanding will be cultivated by the reasoning pursued to deal with disagreement. Moreover, because conflict within collaboration can highlight the existence of differing perspectives, so it can stimulate a range of insights regarding the very nature of a rational understanding. For example, it can help break down young people’s tendency to see causality in ‘one-variable’ terms – complex circumstances understood has having singular causes (Grotzer, Derbiszewska and Solis, 2017). Collaborative talk can reveal that a partner holds a contrasting perspective to one’s own, yet the two perspectives can seem equally valid and perhaps compatible. Collaborating allows the exploration of causality around such complementary perspectives, thereby promoting a stronger sense of how phenomena can arise from multi-variate influence. Moreover, such occasions that require the management of differing perspectives can prompt a wider understanding that knowledge must invariably be a matter of human judgement (Kuhn, 2020).
It may feel uncomfortable to organise conversations that welcome conflict. The same uneasiness may be felt for the encouragement of argument as a format for problem-solving. Yet arguing is a powerful arena in which knowledge can be re-configured by collaborating partners (Kuhn, Zillmer et al, 2013). Of course, arguing is by no means an unknown experience for young people. Yet it does not follow that its ‘schooled’ version will be straightforward. Strengthening young people’s confidence in pursuing disciplined argument is difficult. One large-scale study shows that interventions with 11–16-year-olds can be unsuccessful even if extended over two years of schooling and reinforced by extensive teacher development programmes (Osborne, Simon et al, 2013). Indeed, Kuhn and Lerman (2021) have demonstrated in two studies how 13–14-year-olds may manifest only limited appreciation of what argument requires as a productive reasoning process. They conclude: “…the desired understandings of the link between evidence and claim and of the limitations of different forms of evidence in relation to a claim are not understandings that we can assume to be in place in young adolescent students, even ones who have had frequent opportunities to engage in evidence-based science inquiry activities” (p.1048-9). Nevertheless, the researchers suppose that persistence with the schooled organisation of collaborating-as-argument will help cultivate this powerful form of reasoning (Shi, Matos and Kuhn, 2019)
The harmonious management of conflict involves a requirement to be accountable for one’s own assertions. In offering a contribution, the individual must make sense of it to collaborating partners. But, in doing so, they will be making sense of it to themselves. Not only tabling something that they know or believe but also explaining why that knowledge or belief is held by them. In short, a collaboration can create pressure for what psychologists’ term meta-cognition. That is, being reflective about : a self-conscious awareness of what one knows and, thereby, more confidence in its management and application. Outside of collaborations, many students struggle with presenting their ideas along with adequate explanation or expansion. For example, Roscoe and Chi (2008) illustrate how undergraduates when tutoring their peers will often respond to student uncertainty by restating or paraphrasing information, without significant elaboration. It is possible that interacting with a collaborator (rather than a tutored peer) prompts a richer dialogue. The scrutiny and interrogation that is provided by a thinking-together collaborator creates a pressure to confront the grounding of what the individual knows and so become more reflective about the nature of their own knowing.
The coordination of collective thinking and the management of its conflicts are two high-order capabilities that are fostered by schooled collaborations. There is much research documenting the prominence of those responsibilities and some of the strategies adopted at different ages to realise them (Järvenoja, Törmänen et al, 2023). However, collaborating can also benefit from management from external agents: teachers, designers and technologies can play a role optimising the experience of shared thinking. Many important decisions will always lie in the hands of those who are closely familiar with the characteristics and history of potential collaborative partners. Little can be said here by way of generalising for such matters that are so local. However, two themes of a more over-arcing nature that related to such external structuring are briefly considered in the next section.
3.3 Collaborating: Optimising its prospects
Research evidence indicates that collaborating is an effective act of learning (2.2). However, while that is the typical finding, the evidence is mixed. Moreover, research on its classroom application finds that teachers can be uncertain: sometimes positive (e.g. Liebech-Lien and Sjølie, 2020) sometimes hesitant (e.g., Abramczyk and Jurkowski, 2020). Given all that has been said above regarding the psychological dynamic of collaborating, what can be suggested by way of optimising the prospects of collaborating as an act of learning?
The discussion in 3.2 above highlights the uncertainty of outcome that is associated with differences in participation. The collective nature of thinking in a collaboration can marginalise some of its members. Where this happens, those individuals will be denied the important generative thinking that collaborating invites. The less engaged participant may be a witness of the group’s conclusions but not be an active agent in reaching it. However, failing to engage deeply with the generativity of group thinking may arise from its members not reliably articulating that generative progression. Contributions are tabled but not publicly derived. The value of making reasoning fully visible in that sense is a responsibility that can be cultivated. At least three routes towards this can be suggested: strategically convening group membership, pre-training members in the necessary sensitivity, and inserting external prompts to guide the discourse.
First, some commentators have suggested that the ‘social pedagogic potential’ of a classroom community gets neglected when organising group work (Blatchford, Kutnick et al 2003). For example, intuition suggests that pre-existing friendships might encourage members to be more animated in sharing the course of their reasoning. Certainly, students will often profess a preference for working with friends (Lintner, Diviák and Nekardova, 2024). While Kutnick and Kington (2005) have reported one study involving years 1, 3 and 5 of primary school that explores friendship pairings in science reasoning tasks. Such pairings were found to facilitate performance but they could also inhibit it. In particular, girls’ friendship pairings has the highest performance while boys the lowest. It is no surprise that a history of mutual familiarity will be brought to the demands of classroom collaborating. However, taking advantage of it is unlikely to be guided by simply rules but depend on the insight of an overseeing teacher.
Second, individuals will come to a collaboration with different levels of experience and confidence around that form of communication. An influential concept in psychology for representing this kind of social or cultural knowledge is the ‘script’. The textbook illustration of this is the restaurant script. Cumulative experience of visiting restaurants equips us with an understanding of how to act and what to expect of others. By analogy we may acquire scripts for a collaboration or scripts for forms of collaboration – such as an argument. Researchers have worked at cultivating within classrooms the discourses of social reasoning (Mercer, Dawes et al, 2004). Pre-establishing greater confidence with the scripting of collaborative talk can make a difference to the quality of shared problem solving (Mercer and Littleton, 2007). However, such gains may not be successful when pursued as short-term interventions. Jurkowski, Mundelsee, and Hänze (2024) implemented a ninth-grade programme to foster transactive discourse (of referring to, elaborating, and building on the learning partners’ contributions). This advanced the quality of collaborative talk but not sufficiently to bring about changes in learning outcome.
Third, effective discourse could be could be stimulated by incorporating in group work features that explicitly direct it. Intervention studies of the kind mentioned above may be said to be working on students’ ‘internal scripts’ for collaborating. However, the scripting can be external: that is, achieved by prompting conversational moves at key moments (Koller, Fischer and Slotta, 2007). Such insertions can prompt very general moves (“ask for evidence”, “sum up where you are”) or moves that were specific to a particular task, or the point reached in a task. External scripting is a method that has enjoyed some success. Dillenbourgh (2007) has both pioneered this method but also warned caution in its use. He warns against ‘overscripting’. The sometimes-mechanical interventions of a script can disturb the natural flow of interactions, or the natural order in which components of a task are tackled. Moreover, they may increase rather than decrease the cognitive load involved analysis and coordination.
A significant motive for interventions of the sort outlined above is to encourage collaborators into making their reasoning explicit and visible to partners. This maximises the involvement of their partners in the generative nature of the group’s conclusions and, ultimately, promises effective learning its participants. However, the inclusive accessibility of collective thinking may also depend on the environment in which that thinking takes place. For example, although participants may make their reasoning transparent through what they say, the goal of transparency may also be approached by the generating and sharing written notes. Such notes would be an external resource mediating collective progress in the task being addressed. Yet this example raises a very general point about the importance of mediating reference points in the environment of a groups’ thinking. Some problems are not grounded in that way: they will therefore involve a struggle with elements that are abstract and only made visible in speech. This is unlike the LEGO building of our case ‘C’ example collaboration in 2.1. There, an emerging design speaks for itself. Both the history of collaborators’ building decisions and where they might go next are visible in the physical form before them. The example illustrates how the application of intersubjective communication is facilitated by being able to direct partners attention to shared references points in their environment. Indeed, one reason that computers became very central to the cultivation of classroom collaboration may have been their ability to make otherwise abstract components of reasoning visible and shared on their screens (Crook, 2000).
Psychological theory cannot prescribe watertight strategies for the implementation of successful collaborative learning episodes. However, in scrutinising what takes place in such episodes it has derived a set of concepts that may nevertheless inform design decisions that have to be sensitive to features of the local teaching and learning ecology. The examples in the present section illustrate how those concepts can underpin such practical decisions.
4 Becoming collaborative: the collaborative disposition
In this final section, consideration is given to how individuals acquire characteristics of communication and cognition that are favourable to collaborative thinking. That has been termed here a ‘collaborative disposition’. While we feel we may readily enough recognise that disposition in people we know, it is not well-developed as a distinct psychological construct of individual difference. However, what we can do is to highlight the grounding of such a disposition as a commonality within the human condition. The possible value of this perspective is that it can give depth to the confidence we may claim in collaborating as an act of human learning.
Many developmental psychologists believe that from very early in life children manifest social and cognitive skills that are species specific. If so, such findings might well influence how we prioritise collaborating as a practice for learning. A precocious social nature during human infancy must surely imply a ‘preparation’: a natural readiness for learning by seeking engagement with others (Murray, De Pascalis et al, 2016), To make such claims concrete, consider the following two examples of the human child’s deeply social nature. First a study by Hermann, Call et al (2007) in which a range of cognitive tests were given to two-year old children. The same tests were also given to a group of our closest primate relatives (chimpanzees and orangutans). Children and apes performed similarly on tests relating to the physical world, but children showed greater skill for tests relating to the social world. Second, and pursuing a similar contrast, Waneken, Chen and Tomasllo (2006) observed the behaviour of children (18-24 months) during play with an adult. Then they observed chimpanzees similarly playing with the same human adult. The children were receptive to both problem solving and social games, but the chimpanzees were uninterested in the social games. Moreover, if the adult ceased playing, children attempted to re-start the interaction, while no chimpanzee attempted re-engagement. Where does this begin for us?
4.1 Becoming collaborative: The social newborn
Childhood social precocity is first evident in how newborn children orient to their immediate environment. Simion and Giorgio (2015) have reviewed a long line of research demonstrating a neonatal bias for visually attending to human faces. They concluded that this reflects a pre-set preference for a structural visual pattern whose configural properties are intrinsic to faces (i.e., up-down asymmetry of component parts and the congruency of those inner features). Similarly, the newborn infant shows preferential attention to human speech and will orient to maternal voices in preference to those of female strangers (DeCasper and Fifer, 1980). For sure, this is not collaborating, but it is a helpful start.
Then, across the first year, children’s development progresses towards behavioural attunements with key people (Stern, 1971). At first, it occurs as simple behavioural imitation (McGowan and Delafield-Butt, 2022; Malatesta and Haviland, 1982). Although these are exchanges with emotional potency: something well illustrated by the ‘still face procedure’ (Tronick, Als et al, 1978; Mesman, van IJzendoorn et al, 2009). These researchers arrange for a routine adult-infant interaction to be interrupted by the adult unexpectedly dropping into a still face. Infants quickly look away and show loss of positive affect. Such exchanges illustrate infants adjusting to the behaviour of another individual by using feedback from within their interaction. This illustrates our very early involvement with the rule-led nature of social exchange.
Lavelli and Fogel (2005) document how these patterns in the first three months cause a shift from simple attending to the world (i.e., with limited emotional expression), towards an increased attention-with-emotion and, thereby, the communicative expression of our emotional states (Cole and Moore, 2015). Three months is often cited as a landmark. Two particular developments come together. First, more complex playful agency and, second, exchanges with others that are increasingly integrated, continuous, and sustained. Synchronies that Bateson (1979) termed infant/adult ‘proto-conversations’ start to form into sequences that – by extension from ‘conversation’ – might be regarded as ‘proto-narratives’ (McGowan and Delafield-Butt, 2022). A ‘reach-to-touch’ develops into a ‘reach-to-grasp’. A ‘reach-to-grasp’ develops into a ‘reach-to-grasp-to-drink’, and so on. Receptive adults will provide partners for such narratives to evolve with further complexity and reach.
4.2 Becoming collaborative: the foundation of Intersubjectivity
The nature of human intersubjectivity has been outlined earlier in this chapter (3.2). In relation to human infancy, it is common to distinguish primary and secondary intersubjectivity (Trevarthan and Aitken 2001; Terrace, Bigelow and Beebe, 2022). In the primary form the infant’s subjectivity matches with a partner: they simply recognise and display complementary emotional states. While the secondary form (typically occurring around 9 months) integrates the subjectivities of both infant and partner with some third-party presence. In this ‘secondary intersubjectivity’ the infant and partner achieve mutual awareness with respect to an external state of affairs (an artefact or event) that exists ‘between’ them. For example, perception becomes jointly focussed on a toy in their shared space. Such mutuality of understanding then allows progress towards deliberate shared action: a proto-collaboration.
The ‘mindreading’ capability of human children has been most intensively researched for periods later than infancy – most often around 4-5 years. At that time, children come to understand that other people have desires and beliefs (Beaudoin, Leblanc et al, 2020). However, Tomasello and colleagues have argued that the foundational skill for human sociality – and the grounding for collaboration – is not a desire/belief psychology (although that becomes important later), but a more basic understanding of intention in others (Tomasello, 2020; Tomasello and Gozalez-Cabrere, 2017). They propose that human development is propelled by an integration of this mindreading capacity with all the various opportunities for interacting with others as are provided by a child’s local cultural context. It is this integration that defines the real roots of a human capability for (and appetite for) collaboratively thinking together.
This brief review of the roots for human sociality suggests two summarising propositions. First, that human infants are inherently prepared for interacting with others and, second, that such preparation serves to resource our inevitably collaborative nature. With regard to the first: cross-cultural research is needed to test the claimed universality of precocious sociality. After all, so many of the influential studies cited above have been conducted in the developed economies of the West. It is established that infant experience can vary widely as a function of culturally specific child-rearing practices (Halberstad and Lozada, 2011; Lancy, 2008; Lancy and Grove, 2010; Weisner, 2002). Unfortunately, research clarifying the consequences of this variation is scarce. However, whenever the universality of findings for Western contexts are challenged, it is usually in relation to those child-rearing practices elsewhere that describe adults as more passive in their relationship with infants (Levine, Dixon et al, 1994). Yet close study of those ecologies does suggest that sensitivity still exists (and adult/child synchrony still does follow) but it is concentrated in very particular (and sometimes brief) occasions of physical contact routines (e.g., Mesman, Basweli and Misali, 2021). Similarly, child-rearing practices involving rare face-to-face interaction may be contexts where those principles of interaction discussed above are exercised more through touch than the visual modality (e.g., Wefers, Schuhmacher et al, 2023). So, the first conclusion for this Section may be relatively secure: it appears that human infants do come well prepared for social interaction.
The second of the two summarising propositions might then be that human beings are inherently collaborative. Yet if there is a feature ‘inherent’ to being human it is not so much collaboration as intersubjectivity. Certainly, the human capability for mindreading can be mobilised for purposes of collaborating. And it surely is. But it can also be mobilised for competition. After all, co-labour can even be both harmonious and tense within the same project or episode. Indeed, the interpersonal dynamic can be sufficiently volatile, that the success of some arranged collaborations often cannot be ensured. For instance, partners for a classroom collaboration may arrive with very different ‘intersubjective attitudes’. Variation perhaps determined by the appeal of the task set, or by a history of earlier social interactions, or just by a passing mood. Such variation could mean resistance, more than harmony. So, the second conclusion for this Section must be that human children are inherently prepared for collaborating. But contextual circumstances will determine whether and how that preparation is enacted. Put another way, intersubjectivity is not a promise that collaborating will always and everywhere drive learning.
Nevertheless, the social orientations and interactions of infancy create the foundations for collaborating to become a significant act of learning – subject to favourable conditions. This means asking next how such a foundational readiness is articulated within the subsequent scprehool period: experiences that will prepare children for the more structured peer working often required by schooling.
4.3 Becoming collaborative: the preschool years
Numerous research reports critically observe children acting together in the post-infancy years, seeking signs of collaborative motives, competencies, and attitudes. The striking sociality of infancy implies peer collaboration will emerge and flourish during this pre-school period. After all, it was noted above that by the end of their first year, infants can read the intentions of others around them – an ability foundational to collaborating. However, although 1-year-olds can interpret intention within the ongoing actions of adults, they are unable to read an adult’s (or peer’s) apparent intention in advance of those actions (Meltzoff, 1995).
In observing young children talking together during free play, it is easy to worry about whether what is being observed is or isn’t ‘genuine’ collaborating. However, this may distract us from a more significant topic: namely, understanding the trajectory of early thinking-together, a journey that starts in the early preschool years. Where that journey will take children is towards engagement with an increasingly formalised version of the playful exchanges sketched above. A version that was termed in this chapter ‘schooled collaborations’. Evidence reviewed by Ramani and Brownell (2014), shows that joint talk within the free play of pre-schoolers demonstrates many features that could be mobilised later within schooled collaborations. Sadly, there is little research exploring how individual differences in such early joint thinking correlates with achievements later in those more structured collaborations-to-learn. Certainly, pre-schoolers do demonstrate skills of coordinating their goals and intentions with others during play, but even at the end of the pre-school period they do not settle easily into schooled collaborations – when these are offered in advance of formal school experience.
A good example of children on this collaborating trajectory is from a study by Ashley and Tomasello (1998). They invited four sets of dyads aged 24, 30, 36 and 42 months to work a lever-and-tubes instrument that, through shared effort, could release a desired object. Even when adults were assisting, children of 24 months never managed the task together. Children of 36 months became more proficient but were still slow to cope. While the 42-month-old children did solve the task relatively quickly and fluently. This pattern was interpreted in terms of children’s developing understanding of a task partner’s perspective. The 2.5- and 3-year-olds showed some limited sense of peers as having alternative perspectives. While the 4-year-olds seemed much more aware of their partners as actually being complementary actors, with perspectives, ambitions, and beliefs of their own. Research identifies 4-5 years as the age at which intersubjectivity starts to deliver something beyond a sensing of others as having intentions. It delivers an awareness of others also having desires and beliefs behind those intentions – and thus how these states will allow the coordinating of action. Similar findings to those of Ashley and Tomasello have been reported by Brownell, Ramani and Zerwas (2006) and Castellaro and Roselli (2015).
What has been said here about the preschool period can be summarised. There is a distinction to be made that applies to how readily children exercise their thinking together. There is that which happens in the playground and then there is the directed collaborating typical of the more orderly classroom. What happens in the playground gets termed ‘collaborating’ (with some hesitation). However, the communication observed there – however labelled – is a social coordination that will become increasingly structured within the experience of formal schooling. Observations suggest that, by the end of the preschool period, many children are equipped with tactics suitable for such schooled collaboration: tactics more tuned to classroom problem-solving tasks. However, prior to the influence of schooling, when pre-schoolers are prematurely confronted with such classroom-style collaborations, their progress together is halting.
Some researchers are well aware of the tensions that surround adapting to schooled collaboration. So, Ramani and Brownell (2014) argue that any research-led understanding of children’s confidence in ‘thinking together’ (and readiness for schooling) is better derived from observations of free play than from observations of experimenter-managed preschool collaboration tasks. Moreover, some teachers may reason similarly. For example, when they set up in-class opportunities for play that are not explicitly structured for learning. Ogden (2000) reports such a project, commenting: “…shared activities between peers provide valuable opportunities for children to engage in collaborative activity in the first years of schooling; such opportunities allow them to develop and explore their roles as collaborators” (p.224, emphasis added). Thereby children are brought together in a way that hopefully ignites more structured thinking together through play. In her reflection on this, Ogden is articulating a perspective that is quite widely shared: namely, that experience of making an effort to think together is worthwhile as an end in itself.
5: Chapter summary
In common with other chapters in this book, this final section highlights the key psychological foundations that have been identified for the present act of learning. If this chapter is longer than others it is partly because the topic has required introducing a significant number of those foundational ideas. Many of them will recur in subsequent chapters where their character will be further reinforced.
Collaborating is a practice of integrating ones thinking with that of other people while pursuing a shared ambition of reaching some agreed goal. This formalising of the familiar act of thinking with others was termed a ‘schooled collaboration’. Learning is a common outcome: one detected by pre- and post- testing knowledge relevant to the goal reached at a collaborative episode. Summaries of research on the efficacy of collaborating conclude that it generally (but not universally) supports learning: when outcomes are statistically compared with those from solo study.
The effort of collective thinking was presented as illustrating the externalising nature of mental life: i.e., the human ability to offload cognition by incorporating resources available in the external environment. In the case of collaborations, those resources are the thinking presence of collaborating partners. Learning from such thinking together has been informed by studying brainstorming groups: they reveal the hidden presence of collaborative inhibition. The inhibitory influence of thinking with others was manifest in different ways. First, as the social disengagement of participants (ranging from elective social loafing to social exclusion), second as the disruption of cognitive retrieval and integration, and third as a vulnerability to premature fixation on ideas arising from other people, sometimes early in discussion. The structural underpinning of these inhibitory factors was illustrated in relation to human long-term memory and working memory. The documentation of these inhibitory forces has not challenged the general finding of efficacy for collaborative learning. But it has highlighted the importance of thinking that is generatively directive towards problem solutions and how that generativity can limit the depth of learning by individuals in a group, or limit its distribution within the group’s composition.
Central to the exercise of a generative (or ‘mutually constructive’) mode of thinking are two important forms of collaborative discourse: coordination and conflict resolution. Coordination can help achieve a convergence of group thinking. It does so by effectively managing the transactions from the interlocking of partners’ thinking towards some consensus. This is an achievement that recruits the intersubjectivity of human learners. The existence of conflict forces pressure on the kind of thinking that highlights and resolves different thinking perspectives. But by requiring individuals to articulate the grounds for their conflicting perspective, that tension can be an occasion for meta-cognition – a greater conscious awareness of one’s own thinking.
Optimising collaborating as an act of learning demands attention to the internal management of generative thinking. That, in turn, depends upon designing collaborations to emphasise the responsibility of participants to articulate their reasoning while being receptive to the reasoning of partners. Three design opportunities were identified: teachers making reference to the local social pedagogy of a classroom, working to develop the internal scripts of collaborators, and imposing external scripts in the form of a structured prompting of discourse.
Finally, the foundational nature of collaborating as an act of learning was emphasised by outlining the emergence of a collaborative disposition in early life and how it evolves into the ‘schooled collaborations’ that have been the central topic of the present chapter.