Why mistakes must be celebrated

I have just read a book by Matthew Syed entitled ‘Black Box Thinking‘ (2015), John Murray.

Despite the title, it is about, learning from mistakes. The ‘black box’ reference is literal. It refers to the black box data in crashed aircraft that provide the evidence needed to learn from mistakes that caused or contributed to the crash. It addresses a vast range of bad event contexts from airline safety through deaths and harm caused by medical errors, the mental contortions of politicians faced with dire negative consequences of their decisions ( eg invading Iraq), the unwillingness of police and prosecutors to accept the innocence of those wrongly convicted of serious crimes when subsequently vindicated by solid evidence (often DNA), to wrong-headed decision making in business, scientific research, sport and more.

It is a powerful articulation of the vital, perhaps universal, role of mistakes in all deep learning.

Syed identifies two cultural attitudes to mistakes. He refers to these as ‘closed loop’ and ‘open loop’ responses to failures. This is not to be confused with the use of the same terms in engineering feedback systems.

Syed’s ‘closed loop’ is where information doesn’t lead to progress because errors and weaknesses together with their consequences are misinterpreted or ignored. An ‘open loop’ is where feedback from a failure is rationally and thoroughly analysed leading to learning that is acted upon.

An example of the former is the historic medical practice of ‘bloodletting’, which frequently led to the death of the patient either directly through the loss of blood or indirectly because the blood loss so weakened the natural systems that the patient succumbed to the true agent causing the illness. Despite continuously poor outcomes, ‘bloodletting’ continued as a medical practice for more than 1500 years.  As late as 1942, a medical textbook considered bloodletting appropriate treatment for pneumonia. The ‘closed loop’ culture resisted any proper analysis of outcomes. If a patient survived the doctor would say, “Boodletting cured him”. If not, the explanation was also clear. The patient must have been extremely ill if the wonder cure of bloodletting did not work.

‘Open loops’, where mistakes are analysed and learning takes place, have long been the modern scientific assumption, but as Syed shows through many examples, especially in healthcare, this often does not take place in practice.

Nature uses ‘open loop’ feedback in response to failure. The clearest and most ancient example is the mechanism of ‘natural selection’ that drives the process of evolution of all life on earth. Failure is the death of individuals prior to reproduction. The feedback from this failure directly drives the changing make up of the gene pool of the population of the species.

Open loop feedback is genetically hard wired into the mechanisms of perception in animals. The responses from sensory nerves are processed by the brain resulting in both conscious images (vision) and subconscious responses such as by the autonomic nervous system: a part of the nervous system that regulates key involuntary functions of the body, including the activity of the heart muscle; the smooth muscles, including the muscles of the intestinal tract; and the glands.

This is well understood in the human sense of vision, where we do not perceive the visual image that accurately reproduces the physical optical image projected by the eye lens onto the retinal nerves. Instead, the brain processes this raw image so as to make the best guesses from optical ambiguities and perceived threats to produce conscious images in the mind that are optimal for the success of the individual. This process relies on algorithms produced by millions of years of the evolutionary feedback from failures. This feedback is somehow processed in brain tissue in ways that are a very long way from being understood.

The achievements of science and technology are the result of the adoption of the ‘scientific method’, which is directly driven by the principle of learning from failure. Scientific understanding is not the celebration of experimental confirmations, but of the failures of attempts at refutation. A new scientific theory results not in a rush for experimental confirmation, but in attempts to refute it, especially by those that believe it to be true. This is the famous principle described by Karl Popper.

Technological achievements are often a combination of scientific theories tested by such Popperian failures and actual physical failures and crashes (literally). The windows in jet aircraft are rounded in shape rather than square as a direct result of the crashes of the pioneering Comet jet airliner in the 1950s. These identified the role of window shape in the development of fatigue cracks.

Feedback from failures can also drive technological innovation through cumulative marginal gains that result from ‘small step’ iteration techniques that work when there is either no well understood theory or else the complexity of the system is too great to be theoretically analysed. Syed gives some powerful examples of this in sport, design and business.

Part 2 of ‘Black Box Thinking’ is about ‘Cognitive Dissonance‘, which is one the main mechanisms of not just failure to learn from mistakes, but in explaining why decent professionals so often adopt defensive, cover-up, self delusional attitudes when they make mistakes. Cognitive Dissonance is almost always a factor in human ‘closed loop’ responses to mistakes and failures.

Cognitive Dissonance (inner mental conflict) is a universal human psychological response that mitigates against open disclosure and analysis of one’s own mistakes and failures in order, consciously and/or subconsciously, to protect self esteem and reputation. Syed shows that the higher the status of the person involved, the greater the degree of personal cognitive dissonance and the greater the self-delusion and public obfuscation. He quotes by way of example the public statements about Saddam Hussein’s Weapons of Mass Destruction (WMD) by the British Prime Minister Tony Blair prior to and after the invasion of Iraq.

Syed only briefly addresses the issue of ‘cognitive dissonance’ in school-based learning, which is the subject of my book , ‘Learning Matters‘. In it I describe the ‘closed loop’, common sense learning theory of ‘behaviourism’ in contrast to the counter-intuitive ‘open loop’ learning theories of ‘developmentalism’. This is also explained on my website.

Syed’s approach has further informed my thinking in terms of the universality of ‘learning from failure’. My book, Learning Matters sets out an alarming analysis of the marketised English education system in which I argue that the result of competition between schools is driving a destructive form of market choice that reduces educational standards, while at the same time securing a ‘closed loop’ national justification in the form of continuously rising exam results that hide the resulting decline in deep understanding and failure of development of the cognitive function of our learners. This is further addressed on my website here and here.

Syed is right to draw attention to the positive features of market forces in which competition and market failures drive technological advances and product improvements.

However I argue that marketisation in the area of public services is more likely to result in the degrading of quality through the emergence of flawed performance indicators that drive corrupting perverse incentives. This is fully explained in my book.

A direct mechanism is the imposition onto schools by the market schools of ‘closed loop’ behaviourist teaching methods adopted by ‘business culture’ school leaders. These are then forced  onto teachers and then by teachers onto their pupils. Such approaches emerge because market choice requires ‘performance indicators’. In the case of consumer products and services these are often obvious and simple to identify and to quantify. In public services this is usually not the case because the ‘performance indicators’ are not selected by voluntary customers, but imposed by governments upon the public that needs and has already paid for the service through taxes. It is the distinction between customers; and patients (state provided hospital services), school students (state provided education) and passengers (state provided transport). Imposed ‘performance indicators’ like for example, the proportion of pupils obtaining ‘C’ grade GCSE passes in a school, always generate unforeseen consequences, that often turn out to be seriously degrading of quality.

In Part 5 of ‘Learning Matters’ I give several examples of ‘Open Loop’ developmental approaches to teaching and learning. These are being displaced from our schools by the pedagogy that fits the ideology of artificially imposed public service marketisation. This is called the Global Educational Reform Movement (GERM), an international ‘closed loop’ educational recipe that originated in the US and which has been adopted in England (but not other parts of the UK) and uncritically accepted by the British media. Most regrettably of all, GERM ideas, which do not stand up to evidence-based analysis, have been taken up not just by the Republicans in the US and the Conservatives in the UK, but also by the Democratic Party in the US and the Labour and Liberal Democratic Parties in the UK. Sweden is currently recovering from the failure of their adoption. The current phase of the English education system is finally being informed by the failure of the system of imposed marketisation to raise real educational standards rather than the illusion created by the inflation of the artificially imposed performance indicators.

I am a retired headteacher and a career-long teacher of science. Most teachers are able to sail through their careers without much concern for, or knowledge of learning theory at all. This has never been a good thing. It was also true for me until I was lucky enough to be seconded full time by the Leicestershire LEA for a year onto the Master of Educational Studies Course (M.Ed.) at Leicester University in 1981. Here, alongside studying theories of learning, assessment and evaluation and the beginnings of the applications of computers in education, I carried out some educational research which was heavily influenced by the work of former Leicester University postgraduate students Michael Shayer and Philip Adey.

These educationalists progressed to Kings College, London where they set up and developed a programme of teaching for enhancing cognitive development based mainly on the learning theories of Jean Piaget. My M.Ed dissertation was a piece of research on the practicability of matching the cognitive demand of the curriculum to the level of cognitive development of individual pupils. This is a much more sophisticated issue than streaming or setting. The basis of my work was Shayer and Adey’s, ‘Towards a Science of Science Teaching’ (1981). In Learning Intelligence (2002), Shayer and Adey set out six principles on which their Cognitive Acceleration pedagogy is based. This became Cognitive Acceleration in Science (CASE) and later Cognitive Acceleration in Maths (CAME). The following is my understanding of the principles.

Schema (plural Schemata) Theory

Schemata are general ways of thinking that apply in many parallel contexts. For example, Inhelder and Piaget (1958) famously identified the schemata of ‘Formal Operations’ (e.g. control of variables and formal mental modelling) that differed from the schemata characterising the previous developmental stage of ‘Concrete Operations’ (e.g. simple classification and various conservations). In maths the schemata of ‘Formal Operations’ might be transposed to include ratio and proportion, and ability to think in terms of variables rather than just numbers of real objects’. Recognition of underlying schemata is necessary in order to construct appropriate learning and assessment scenarios.

Concrete Preparation

Like buildings, developmental learning needs concrete foundations. Given that the objective is to facilitate progression to formal thinking it must be expected that individual pupils are likely to be at different cognitive stages. If the ground is to be successfully prepared for what is later to be presented then this must be founded on common language and examples from the concrete stage that all can share and understand.

Cognitive Conflict (Dissonance)

Exposing pupils to cognitive conflict is central to all teaching for cognitive development. It provides an approach to learning through the deliberate creation of cognitive dissonance in the minds of students followed by supporting them in resolving this dissonance through open peer to peer debate in the face of the evidence they as students have directly experienced from practical lessons designed by their teachers for the purpose.

Requiring students to confront themselves with factual evidence that doesn’t consciously or unconsciously make initial sense to them creates a state of discomforting mental tension. This is the same as Matthew Syed’s ‘Cognitive Dissonance.  But instead of this dissonance arising spontaneously from undesirable mistakes or failures, in this educational setting the dissonance is deliberately created. However, in both cases the goal should be to achieve deep learning through rational blame-free resolution of the dissonance. Also in both cases peer to peer interaction is key.

Syed touches on this in Part 6.14 of his book.

In order for a cognitive conflict to be resolved within the mind of the individual learner then a personal conceptual breakthrough is often necessary. Cognitive development (growth of general intelligence) arises from the accumulation of such conceptual breakthroughs, each of which is like mounting a cognitive staircase.

If the cognitive conflict is too great then the learner might ‘close down’ and withdraw co-operation. This could be at a conscious or subconscious level. In education, hostility to the whole subject area is therefore a possible consequence. This is a commonly reported reaction to students of science and maths when subjected to ‘closed loop’ teaching methods.  Highly skilled teaching and managing of learning is essential to avoid such an outcome.

If there is insufficient cognitive conflict then the learner will just assimilate experiences at a shallow level and there will be no conceptual or cognitive gain. Just as a fundamental change of culture is needed in a school for ‘open loop’ learning to work, the same is true for (say) a hospital setting if the learning needed to enhance patient safety is to take place.

The work of the Russian learning theorist Vygotsky provides a structure to help the teacher plan learning, through his ‘Zone of Proximal Development’ (ZPD) (1978). The ZPD is the level of cognitive challenge beyond which the learner cannot manage unaided, but not beyond what can be understood with the assistance of a teacher or more able peers. The teacher and peer group members can assist in a variety of ways that involve discussion (peers) and skillfully constructed leading questions (teacher). This is a key role of the teacher. It is only by experiencing this type of teaching and subsequently discussing it in departmental teams that the necessary teaching expertise can be built within a school. The same principle would apply to the clinical staff in a hospital.

Social Construction

Vygotsky (1978) asserts that understanding first appears in the social space that learners share. Only then does it become internalised by individuals. Piaget shares this view but expresses it differently. The key point is that in a school the resolution of cognitive conflict in an individual learner needs to be a social process. The participants can then assist each other in grappling with the cognitive conflict. There is then a shared participation in recognising the resulting cognitive dissonance, which the students are encouraged to welcome rather than fear.

This approach requires a certain quality in the social relationships in the classroom. I was lucky enough early in my career to work in a school where such quality relationships existed, and it became a career goal to eventually achieve it in my headship school. Students have to trust each other and not fear the teacher if they are to risk revealing the true depth of their misunderstandings and the sources of their cognitive dissonance. University Challenge question master Jeremy Paxman famously might need some practise in this. Peer relationships have to be good enough for all group members to be comfortable with revealing their lack of understanding to each other as well as both collectively and individually to the teacher. This is a big ask, not to be underestimated.

A school culture of behaviourist authoritarianism will not provide sufficient personal psychological security for the teacher or the pupils for the necessary social interactions to ignite. Syed addresses this dilemma of the ‘blame free’ approach in the context of the reporting of medical errors in hospitals. Obviously nothing can be achieved where informality degenerates into anarchy and chaos reigns, either in hospital wards or in classrooms, but in practice, where it has been tried in both contexts the outcomes are overwhelmingly positive.

This is a developing field that Syed has researched in healthcare systems. I have researched it in schools and I am excited by the convergence. The work of Vygotsky in relation to education is many decades old. However his ideas inform all contexts where learning is first located and realised on a social plane. Vygotsky informs the process whereby ‘social learning’ becomes internalised by the key individuals involved. This is as true for hospitals, where the key players in the learning milieu are clinicians as it is for schools where the key players are  students and their teachers. Is it time for the proponents of a ‘blame free reporting culture’ in hospitals to discover Vygotsky?

In schools great demands are placed on the teacher, who needs the skills to be able to promote positive self-sustaining, reasoned discussion between pupils. This is very difficult to achieve and the true hallmark of a good teacher. There are regrettably a growing number of schools, many judged as successful in league table terms, where relationships between school managers and teachers, teachers and pupils and pupils with peers are so limited in depth and quality that this type of high level ‘open loop’ learning is likely to be impossible.

A blame-free ‘open loop’ learning culture in healthcare poses parallel challenges for managers. Syed describes how formal hierarchical structures with all their authoritarian trappings have to give way to more ‘hands on’ collegial approaches of shared responsibility. My book explains why this has to happen in schools too, and bewails that marketisation is pushing school cultures in the opposite direction.


This means being aware of your own thinking process. It is a personal mental habit essential for the resolution of cognitive dissonance.  It recognises that language provides the tools for thought and that learners benefit by silently but consciously ‘talking to themselves’ as well as talking to peers and the teacher. As learners experience the resulting cognitive development they develop a metacognitive ability that can be characterised in the education context as a higher general cognitive ability (the student gains in general intelligence).

This is probably also true for doctors confronted by their own medical errors and those of others. It is certainly likely that such metacognition makes them both better doctors and more advanced medical professionals.

If the reader is experienced in other disciplines and failure contexts then the parallels should be apparent.

Metacognition is always essential for the energy and tensions of cognitive dissonance to transmute into deep learning. Healthcare professionals need to take on metacognition.


This is applying new understanding to other contexts. It can be called ‘lateral thinking’. Students need to be encouraged to see links with other subjects and disciplines. The teaching methods that are taking over our schools and which teachers are now increasingly compelled to adopt within the league table driven, competitive English education system, are ‘closed loop’ approaches that do not result in cognitive growth. If cramming and repetition reinforced by rewards and the shame of failure have any effect at all on individual cognition then it is cognitive depression leading to rejection of challenging concepts and destructive cognitive alienation.

‘Closed loop’ and ‘open loop’ approaches in the education context have their origins in the two classic alternative theoretical models of learning. Shayer and Adey’s ‘open loop’ Cognitive Acceleration is rooted within the ‘developmental’ school that emphasises the importance of challenging and refining the individual mental models (schemata) held by learners.

The ‘closed loop’ approach comes directly from the ‘behaviourism’ of B F Skinner, who believed that all learning was acquired as a result of conditioning resulting from rewarding correct responses. Think rats, dogs or pigeons in cages. Despite the fact that the behaviourist position has been largely discredited and abandoned by learning theoreticians, it corresponds more closely to the ‘common sense’ model usually held by the public, and regrettably the politicians in charge of our education system that should know better or at least be better advised.

Note the contrasting patterns. ‘Open loop’ learning feeds on mistakes and failures, that are not punished, but which are recognised as essential to the development of understanding. ‘Closed loop’ learning feeds on the rewards and ‘reinforcement’ that result from a culture that praises success and ‘getting things right’. Such a culture incentivises negative personal responses to cognitive dissonance, the discouraging of metacognition and it rewards denials and cover-ups. In education the consequence is cheating by the pupil and/or ‘teaching to the test’ strategies, which amount to cheating by the teacher. Both are likely to be overlooked by the business culture senior managers involved (most likely with ‘executive’ job titles).

Note that lessons that ‘cognitively develop’ require pupils to be confronted with cognitive problems ‘just above’ the level at which they can solve them without assistance using their existing mental models (Vygotsky).

Teachers within the ‘closed loop’ behaviorist culture penalise (or humiliate)  children who make mistakes because both punishments and rewards are part of that culture. Learning is much more likely to be based on repetition and practising problems they can already do, rather than risking failure and cognitive dissonance from presenting students with problems they cannot readily solve.

Cognitive Acceleration is just one example of an ‘open loop’ teaching and learning strategy designed to secure cognitive development and Shayer and Adey represent just one such approach, albeit a very important one with proven effectiveness. The success of such teaching creates challenges for the entire curriculum and the way it is assessed and graded through the GCSE.

Shayer and Adey’s Cognitive Acceleration model is based on discrete highly structured lessons. In my headship school we used CASE and found it to be effective not just in raising pupil performance in science, but generally across the curriculum, as claimed.

Cognitively demanding teaching that involves the recognition, nurturing and  normalising of ‘cognitive dissonance’ on the part of pupils was also found to be highly motivating, but only if there is a learning culture where pupils expect to make mistakes and are comfortable with exploring ideas that lead nowhere as well as those that turn out to be productive. I am proposing a definition of effective schools as those where teachers operate in collegiate professional teams, independent of government political interference, focused on securing cognitive development as an essential teaching and learning objective.

By extending Matthew Syed’s thinking into school-based education I am confirming its universal application to learning of all sorts, for organisations as well as individuals, in all contexts.

It seems to me that by exploring the educational context the profound universal truths embedded in Matthew Syed’s book acquire additional power and even wider relevance.

This entry was posted in Blogs, Uncategorized. Bookmark the permalink.

14 Responses to Why mistakes must be celebrated

  1. Pingback: School improvement is reducing social mobility | Roger Titcombe's Learning Matters

  2. Pingback: Teachers are leaving as government falls short on recruitment, NAO finds | ukgovernmentwatch

  3. Pingback: Educational Lysenkoism is blighting the English education system | Roger Titcombe's Learning Matters

  4. Pingback: The evidence for ‘Plastic Intelligence’ and why it matters | Roger Titcombe's Learning Matters

  5. A.R.B says:

    I would love to see this article expanded, perhaps with contributions from more teachers in the field, to include more concrete examples of behaviourism and developmentalism, including in the work-lessons students are given, the qualifications system, Ofsted helicoptering, etc.


  6. Arnold says:

    Interesting article but fatally flawed. Whoever Syed is, doesn’t know his feedback loops at all.
    No creature is hardwired into an “open loop” otherwise eyes, ears, touch, tongue et all would be totally redundant.
    Those senses FEEDBACK. To gain feedback, you need a closed loop. You learn not to touch that nettle, it hurts, not to look at the sun, you see spots, to stick your ear against a wall, you can hear next door. All LEARNED, all using comparative data, that has been processed through your brain, You can then get a sense of loud, bright, do the figures look right, is that hot, so on and so forth.


    • Syed does not use ‘open and closed loops’ in the engineering sense. I make this clear in the article.

      “Syed identifies two cultural attitudes to mistakes. He refers to these as ‘closed loop’ and ‘open loop’ responses to failures. This is not to be confused with the use of the same terms in engineering feedback systems.”


  7. Pingback: Telling isn’t teaching and listening isn’t learning | Roger Titcombe's Learning Matters

  8. Pingback: Is effective teaching taking place in Academies? | Roger Titcombe's Learning Matters

  9. Pingback: Demonstrating the top five EEF Toolkit teaching interventions using the kitchen sink | Roger Titcombe's Learning Matters

  10. Pingback: School Councils – misunderstandings, misuse and missed opportunities for learning excellence | Roger Titcombe's Learning Matters

  11. Pingback: The girl who broke into lessons | Roger Titcombe's Learning Matters

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s