Tuesday, 31 December 2019

Programming the Educational Platform: A turning point for educational technology

The sale of Instructure to a private equity firm, Thoma Bravo, has prompted various reactions within education (see for example https://eliterate.us/instructure-plans-to-expand-beyond-canvas-lms-into-machine-learning-and-ai/). Instructure's Canvas has established itself as the leading VLE, with many high-profile institutions opting for its clean interface, seamless video and mobile integration, and powerful opensource service-oriented architecture. It is ahead of the competition, correctly identifying the move towards data-oriented educational coordination and flexibility, and providing an impressive range of tools to manage this.

The indications that there has been some serious thought behind the platform include its GraphIQL query interface (see https://github.com/graphql/graphiql), and an API which sits beneath Canvas's own interface. The API is surprisingly easy to use: simply adjust almost any Canvas URL for pages, files or modules to include "/api/v1/" after the domain, and instead of the interface, you get a JSON file. The consistency of this is impressive, and putting data in (automatically creating content, for example) is as easy as getting data out.

Instructure, like many players in educational technology, see their future in data (Turnitin was also sold this year for £1.3 billion). Of course, like Turnitin, in providing a hosted platform, they have access potentially to enormous amounts of data. The big hope for the corporations is machine learning and predictive analytics. However, for all the hand-wringing going on, I think it would be wise to be slightly sceptical about what has been portrayed as a corporate data-grab of universities. After all, machine learning is in its infancy, and there is no evidence as to what might be learnt through analysing VLE data that would be useful for educational institutions. MOOC data, after all, was something of a disappointment.

Of course, people point to Facebook, Google and Amazon as corporations which exploit the data of their users. Logic would suggest that education would follow the same path. But the difference lies in the fact that Facebook, Google and Amazon are all trying to sell us things (which we usually don't want), or get us to vote for people (who may not be good for us).

Despite all the claims around the marketisation of education, education is fundamentally about relationships, not sales. So Instructure might be wrong. We should use their tools and watch the space patiently - but I doubt that we are looking at an educational equivalent of Blackrock (although I'm sure that is what Instructure envisage)

The approach of Canvas to rationally organising the technical infrastructure of institutional learning systems is a good one and much needed. Whatever challenges educational institutions face in the future, they are likely to need to adapt quickly to a fast changing environment and increasing complexities (more students, increasing diversity of learning needs, more flexibility in the curriculum, more work-based education, etc). Rigid technical infrastructure which limits control to manipulation of poor interfaces, hides data, and makes coordination difficult will impair the institution's ability to adapt. Instructure has addressed many of these problems. So, basically, the technology is very good - this is what institutions need right now (I'm sure other VLE providers will learn from this, but at the moment they seem to be behind).

This also spells an important change for those whose role is to coordinate learning technology. Data analysis and effective control (probably through programming interfaces) are going to become essential skills. It is through these skills that flexibility is maintained. As more and more content abounds on the internet freely, as video production tools are available to everyone (including students), as the creativity and variety of expression and production becomes more important for personal growth, the job shifts to managing the means of coordination, rather than the production of yet more content. The challenge is for each institution to take control of its own platform - and this will demand new skillsets.

This is a new stage of educational technology. Where MOOCs provided content, they thought little about coordination and relationships, and the essential role of institutions in managing this. In Coursera and Edx, the institution was merely a calling-card - exploited for its status. In creating a flexible technical framework for institutions, initiatives like Canvas approach educational problems as problems of institutional organisation. There is inevitably a trade-off between big corporations which provide the resources to refine these kinds of tools, and institutional needs which when correctly analysed can use them profitably.

The interesting thing about where we are is that both universities and technology corporations are organic entities which swallow-up their environments. In biological terms, they could be said to be endosymbiotic. Lynne Margulis's endosymbiosis theory described how competing entities like this (in her case it was cells and bacteria) eventually learn to cooperate. Is this what we're going to see in education? If it is, then I think we are at a turning point.

Sunday, 29 December 2019

From 2d to 3d Information Theory

I've been doing some work on Shannon information theory in collaboration with friends, and wrote a simple program to explore Shannon's idea of mutual information. Mutual information is the measurement of the extent to which two sources of information share something in common. It can be considered as an index of the extent that information source A can predict the messages produced by information source B. If the Shannon information of source A is H and the Shannon information of B is Hb, then the mutual information is calculated by:
H + Hb - Hab  
There is an enormous literature about this, because mutual information is very useful and practical, whilst also presenting some interesting philosophical questions. For example, it seems to be closely related to Vygotsky's idea of "zone of proximal development" (closing the ZPD = increasing mutual information while also increasing complexity in the messages).

There are problems with Mutual Information. With 3 information sources, its value oscillates between a positive and negative value. What does a negative value indicate? Well, it might indicate that there is mutual redundancy rather than mutual information - so the three systems are generating constraints between them (see https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3030525)

Negative values should not occur in two dimensions. But they do. Loet has put my program on his website, and it's easy to see how a negative value for mutual information can be produced: https://www.leydesdorff.net/transmission

It presents two text boxes. Entering a string of characters in each immediately calculates the entropy (Shannon's H), and the mutual information between the two boxes.


This is fine. But if one of the information sources has zero entropy (which it would if it has no variety), we get a negative value.

So what does this mean? Does it mean that if two systems do not communicate, they generate redundancy? Intuitively I think that might be true. In teaching, for example, with a student who does not want to engage, the teacher and the student will often retreat into generating patterns of behaviour. At some point sufficient redundancy is generated so that a "connection" is made. This is borne out in my program, where more "y"s can be added to the second text, leaving the entropy at 0, but increasing the mutual information: 

But maybe I'm reading too much into it. It seems that it is a mathematical idiosyncrasy - something weird with probability theory (which Shannon depends on) or the use of logs (which he got from Boltzmann). 

Adding redundant symbols is not the same as "adding nothing" - it is another symbol - even if it's zero. 

The bottom line is that Shannon does not have a way of accounting for "nothing". How could it?

This is where I turn to my friend Peter Rowlands and his nilpotent quantum mechanics which exploits quaternions and Clifford Algebra to express nothing in a 3-dimensional context. It's the 3d-ness of quaternions which is really interesting: Hamilton realised that only quaternions could express 3 dimensions.

I don't know what a quaternionic information theory might look like, but it does seem that our understanding of information is 2-dimensional, and that this 2-d information is throwing up inconsistencies when we move into higher dimensions, or try weird things with redundancy.

The turn from 2d representation to 3d representation was one of the turning points of the renaissance. Ghilberti's "Gates of Paradise" represents a moment of artistic realisation about perspective which changed the way representation was thought about forever.

We are at the beginning of our information revolution. But, like medieval art, it may be the case that our representations are currently two-dimensional, where we will need three. Everything will look very different from there.

Tuesday, 24 December 2019

Out of Chaos - A Mathematical Theory of Coherence

One of my highlights of 2019 was the putting together of a what is beginning to look like a mathematical theory of evolutionary biology, with John Torday of UCLA, Peter Rowlands in Liverpool university, using the work Loet Leydesdorff and Daniel Dubois on anticipatory systems. The downside of 2019 has been that things have seemed to fall apart - "all coherence gone" as John Donne put it at the beginning of the scientific revolution (in "An Anatomy of the world"):

And new philosophy calls all in doubt,
The element of fire is quite put out,
The sun is lost, and th'earth, and no man's wit
Can well direct him where to look for it.
And freely men confess that this world's spent,
When in the planets and the firmament
They seek so many new; they see that this
Is crumbled out again to his atomies.
'Tis all in pieces, all coherence gone,
All just supply, and all relation;
Prince, subject, father, son, are things forgot,
For every man alone thinks he hath got
To be a phoenix, and that then can be
None of that kind, of which he is, but he.
The keyword in all of this (and a word which got me into trouble this year because people didn't understand it) is "Coherence". Coherence, fundamentally, is a mathematical idea belonging to fractals and self-referential systems. It is through coherence that systems can anticipate future changes to their environment and adapt appropriately, and the fundamental driver for this capacity is the creation of fractal structures, which by definition, are self-similar at different scales.

In work I've done on music this year with Leydesdorff, this coherent anticipatory model combines both synchronic (structural) and diachronic (time-based) events into a single pattern. This is in line with the physics of David Bohm, but it also coincides with the physics of Peter Rowlands.

When people talk of a "mathematical theory" we tend to think of something deterministic, or calculative. But this is not at all why maths is important (indeed it is a misunderstanding). Maths is important because it is a richly generative product of human consciousness which provides consciousness with tangible natural phenomena upon which its presuppositions can be explored and developed. It is a search for abstract principles which are generative not only of biological or social phenomena, but of our narrative capacities for accounting for them and our empirical faculties for testing them. Consciousness is entangled with evolutionary biology, and logical abstraction is the purest product of consciousness we can conceive. In its most abstract form, an evolutionary biology or a theory of learning must be mathematical, generative and predictive. In other words, we can use maths to explore the fundamental circularity existing between mind and nature, and this circularity extends beyond biology, to phenomena of education, institutional organisation and human relations.

When human organisations, human relations, learning conversations, artworks, stories or architectural spaces "work", they exhibit coherence between their structural and temporal relations with an observer. "Not working" is the label we give to something which manifests itself as incoherent. This coherence is at a deep level: it is fractal in the sense that the pattern expressed by these things are recapitulations of deeper patterns that exist in cells and in atoms.

These fractal patterns exist between the "dancing" variables involved in multiple perceptions - what Alfred Schutz called a "spectrum of vividness" of perception. The dancing of observed variables may have a similar structure to deeper patterns within biology or physics, and data processing can allow some glimpse into what these patterns might look like.

Fractal structures can immediately be seen to exhibit coherence or disorder. Different variables may be tried within the structure to see which displays the deepest coherence. When we look for the "sense" or "meaning" of things, it is a search for those variables, and those models which produce a sense of coherence. It is as true for spiritual practice as it is for practical things like learning (and indeed those things are related).

2019 has been a deeply incoherent year - both for me personally, and for the world. Incoherence is a spur to finding a deeper coherence. I doubt that we will find it by doing more of the same stuff. What is required is a new level of pattern-making, which recapitulates the deeper patterns of existence that will gradually bring things back into order. 

Friday, 20 December 2019

Human Factors and Educational Technology in Institutions

Educational institutions are now enormously complex technological organisations - particularly universities. They are generally so complex that few people in the university really understand how everything fits together. Computer services will have teams who understand individual systems, although it is unusual to find someone in a computer services department who understands how it all fits together technically. Even less likely is it to find someone who understands the divergences of digital practice either in the classroom by teachers, or among professional service staff who process marks (and often organise assignments in the VLE).

Of course, despite the lack of any synoptic view, things keep on going. This works because whatever complexities are created by different systems, an administrative workforce can be summoned up to handle the complexity. Providing marks are entered, exam boards are provided with data, and students progressed through their courses to completion, it might be tempting to ask whether a lack of a synoptic view matters.

This is where technological infrastructure, human factors and organisational effectiveness meet. An effective organisation is one which organises itself to deal with actual demands placed on it. An effective organisation manages its complexity, understands its environment, and has sufficient flexibility to adapt to change.  In a university, it can be very difficult to define "demand" or be clear about "environment". At a superficial level, there is demand from "students" for teaching and assessment. This demand is increasingly framed as a "market". However at a deeper level, there is a demand from society, and the politicians who steer it (and the policy for higher education).  What does society demand of education? In recent years, the answer to that question has also been framed around "the market" - but many commentators have pointed our that this is a false ontology. Society has a habit of turning on institutions which extend their power beyond reasonable limits. There is no reason to suppose this might not happen to universities, which have extended their power through a variety of what Colin Crouch calls "privatised Keynesianism" - individualised debt to pay for institutional aggrandisement such as real-estate (https://onlinelibrary.wiley.com/doi/abs/10.1111/j.1467-856X.2009.00377.x)

Then we should ask "What is the actual demand for learning?" A commonsense answer is that a student should expect to be able to do stuff which they weren't able to do before. But this is a very vague notion. As a crude example, how many of the many thousands of computer science graduates can actually program when they graduate? How many creative writers will make it as authors? How many architects will build a building? Now, of course, this is unfair. There are many "transferrable skills" from a degree - people will go into other walks of life, buoyed by their new-found self-confidence. Apart from those who become so worn-down by assessment and institutional rigidity that their mental health took a knock in education. So there is a first demand: "Education should not leave me feeling worse about myself than when I started".

It turns out to be surprisingly hard to organise that (https://www.hrmagazine.co.uk/article-details/students-regret-going-to-university) Teachers do their best, but within constraints which they and everyone else in the system find unfathomable. Today, many of those constraints are technical in origin. The computer has swamped the natural human pattern of communication which kept institutions viable for centuries. The space for rich intersubjective engagement - whether it is between teachers and students, or between staff, or even between students and their peers - has been attenuated to lights on a screen and clicks and flicks. And with the problems that this creates, the answer is always the same - more technology.

So do we reach the stage where the art of teaching becomes the art of designing a webpage in the VLE? Instructional design would appear to have a lot to answer for. Deep human virtues of knowledge, openness, generosity and revealing of uncertainty do not fit the digital infrastructure and get unrewarded. Flashy new tech sells staff with ambitions and a latent drive for everyone to "do it their way". Some of these people become senior managers and appoint more like them. It's positive feedback.

The equations are simple. All technology creates more complexity in the guise of offering a "solution" to a problem created by some prior technology. Human beings struggle to deal with the complexity, and demand new technical solutions, which often make things worse. How do we get out of this?

The importance of a synoptic view is that it must entail clear distinctions about the system, its operations, its demands and its boundaries. As long as we have no clear idea of the problem to which we want to put technology to address in education, we will be condemned to repeat the cycle. So what is the purpose? What's the point?

It's probably very simple: we all die. The older ones will die first. While they are alive they can communicate something of how to survive in the future. Some will do things which speak to future generations after they are dead. But this can only happen if people are able to live in a world that is effectively organised. In a world of ineffective organisation, complexity will proliferate and the intergenerational conversation will die in a torrent of tweets and emails. There is a reason why the ice caps melt, the stock market booms, and the world is full of mad dictators.



Wednesday, 4 December 2019

Institutions, Art and Meaning

Much of what I am reading at the moment - Simondon, Erich Hörl, Stiegler and Luhmann - is leading me to rethinking what an institution is in relation to an "individual". It's like doing a "reverse Thatcher" (which is a good slogan) - there is no such thing as an "individual". There is a continual process of distinction making (and remaking) and transduction by which "institutions" - as biological organisms like you and me - or families, friendship groups, universities, or societies preserve meaning. This is a Copernican shift in perspective, and it is something that I think Luhmann and Simondon saw most clearly, although there are aspects of their accounts which miss important things.

This is a helpful definition, because it seems we live in a time when our institutions don't work very well. Life in them becomes meaningless and alienating. So what's going on?

I think the answer has something to do with technology. Technology did something to institutions, and a hint at an answer is contained in Ross Ashby's aphorism that "any system which categorises throws away information". Those words echo like thunder for me as I'm in the middle of trying to upgrade the learning technology in a massive institution.

So institutions do something with information which preserves meaning. Institutions which lose information risk losing meaning. Thanks to so-called IT systems, most of our institutions from schools to government are losing information.

I've been thinking (again alongside Luhmann) about art and music. A string quartet or an orchestra is an institution, and through their operations, there is little doubt that meaning is preserved. But what is interesting me is that this preservation process is not simply in the current operations of the group - the practice schedule, or performance for example. It is also something to do with history.

Playing Beethoven is to preserve the meaning in Beethoven. And we have a good idea that Beethoven meant his meaning to be preserved: "alle menschen werden brüder" and all that. What is the mechanism for preserving this meaning? A big part of it is notation: a codification of bodily skilled performances to reproduce a historical consciousness.

The art system preserves meaning over a large-scale diachronic period. It seems commonsense to suppose that if the skills to perform were lost, then the process of preserving the meaning would be damaged. Would we lose this stuff? But is this right? What if the skills to perform are lost, but recordings survive? Some information is lost - but is it the technology of recording which loses the information about performance skill, or does the loss of performance skill necessitate recording as a replacement?

In a age of rich media, "performance" takes new forms. There is performance in front of the camera which might end up on social media. There is a kind of performance in the reactions of the audience on Twitter. But is the nuance of "playing Beethoven" (or anything else) lost?

We need a way of accounting for why this "loss" (if it is a loss) is significant for an inability to preserve meaning. Of course, we also need a way of accounting for meaning itself.

So I will have an attempt: meaning is coherence. It is the form something takes which articulates its wholeness. More abstractly, I suspect coherence is an anticipatory system (borrowing this from the biological mathematics of Robert Rosen and Daniel Dubois). It is a kind of hologram which expresses the totality of the form from its beginning to its end in terms of self-similar (fractal) structures.

The act of performing is a process of contributing to the articulation of an anticipatory system. If information is lost in an institution, or an art system, then the articulation of coherence becomes more difficult. This may be partly because what is lost in not-performing is not information, but redundancy and pattern. Coherence is borne through redundancy and pattern. How much redundancy has been lost in the rituals of convivial meetings within our institutions, where now email of "Teams" takes over?

If our lives and our institutions have become less coherent it is because technology has turned everything into information in which contingency and ambiguity is lost. As Simon Critchley argued in his recent "Tragedy, the Greeks and Us", this loss of ambiguity is a serious problem in the modern world, and it can only be resolved, in his view, through the diachronic structures of ritual and drama. We have to re-enchant our institutions.

I think he's right, but I think we can move towards a richer description of this process. Technology is amazing. It is not technology per se which has done this. It is the way we think.









Saturday, 16 November 2019

Maximum Entropy

On discussing Rossini's music with Loet Leydesdorff a couple of weeks ago (after we had been to a great performance of the Barber of Seville), I mentioned the amount of redundancy in the music - the amount of repetition. "That increases the maximum entropy," he said. This has set me thinking, because there is a lot of confusion about entropy, variety, uncertainty and maximum entropy.

First of all, the relationship between redundancy and entropy is one of figure and ground. Entropy, in Shannon's sense, is a measure of the average surprisingness in a message. That surprisingness is partly produced because all messages are created within constraints - whether it is the constraints of grammar on words in a sentence, or the constraints of syntax and spelling in the words themselves. And there are multiple constraints - letters, words, grammar, structure, meaning, etc.

Entropy is easy to calculate. There is a famous formula without which much on the internet wouldn't work.



Of course, there are lots of questions to ask about this formula. Why is the log there, for example? Just to make the numbers smaller? Or to give weight to something (Robert Ulanowicz takes this route when arguing that the log was there in Boltzmann in order to weight the stuff that wasn't there)

Redundancy can be calculated from entropy.. at least theoretically.

Shannon's formula suggests that for any "alphabet", there is a maximum value of entropy. It is called Maximum entropy. If the measured entropy is seen as a number between 0 and the maximum amount of entropy possible, then to calculate the "ground", or the redundancy, we simply calculate the proportion of the measured entropy to the maximum entropy and subtract it from 1.

Now mathematically, if the redundancy increases, then either the amount of information decreases (H) or the maximum entropy (Hmax) increases. If we simply repeat things, then you could argue that the entropy (H) goes down because it becomes less surprising, and therefore R goes up. If by repeating things we generate new possibilities (which is also true in music), then we could say that Hmax goes up.

No composer, and no artist, ever literally repeats something. Everything is varied (the variation form in music being the classic example). Each new variation is an alternative description. Each new variation introduces a new possibilities. So I think it is legitimate to say the maximum entropy increases. This is particularly true of "variation form" in music.

Now, away from music, what do new technologies do? Each of them introduces a new way of doing something. That too must be an increase in the maximum entropy. It's not an increase in entropy itself. So new technologies introduce redundant options which increase maximum entropy.

If maximum entropy is increased, then the complexity of messages also increases - or rather the potential for disorder and surprise. The important point is that in communicating and organising, one has to make a selection. Selection, in this sense, means to reduce the amount of entropy so that against however many options we have, we insist on saying "it's option x". Against the background of increasing maximum entropy, this selection gets harder. This is where "uncertainty" lies: it is the index of the selection problem within an environment of increasing maximum entropy.

However, there is another problem which is more difficult. Shannon's formula for entropy counts an "alphabet" of signals or events like a, b, c, etc. Each has a probability and each is added to the eventual number. Is an increase in the maximum entropy an increase in the alphabet of countable events? Intuitively it feels like it must be. But at what point can a calculation be made when at any point the full alphabet is incomplete?

This is the problem of the non-ergodic nature of life processes. I've attempted a solution to this which examines the relative entropies over time, considering new events as unfolding patterns in these relations. It's a bit simplisitic, but it's a start. The mechanism that seems to drive coherence is able, through the production of redundancies which increase maximum entropy, to construct over time a pattern which serves to make the selection and reduce the entropy to zero. This is wave-like in nature. So the process of increasing maximum entropy which leads to the selection of entropy to zero is followed by another wave, building on the first, but basically doing the same thing.

In the end, everything is zero.

Sunday, 10 November 2019

Design for an Institution: the role of Machine Learning #TheoryEDTechChat

There's an interesting reading group in Cambridge on the theory of educational technology at the moment. Naturally enough, the discussion focuses on the technology, and then it focuses on the agency of those operating the technology. Since the ontology of technology and the ontology of agency are mired in metaphysics, I'm not confident that the effort is going to go anywhere practical - although it is good to see focus on Simondon, and the particularly brilliant Yuk Hui.

But that raises the question: What is the thing to focus on if we want to get practical (i.e. make education better!)? I don't think it's technology or agency. I think it's institutions - we never really talk about institutions! And yet all our talk is framed by institutions, institutions pay us (most of us), and institutions determine that it is (notionally) part of our job to think about a theory of educational technology. But what's an institution? And what has technology done to them?

It is at this point that my theoretical focus shifts from the likes of Simondon, Heidegger, and co (great though I think this work is), to Luhmann, Stafford Beer, Leydesdorff, von Foerster, Ashby and Pask.

Luhmann is a good place to start. What's an institution? It is a autopoietic system which maintains codes of communication. "Autopoietic" in this sense means that codes of communication are reproduced by people ("psychic systems"), but that the "agency" of people in communicating is driven by the autopoietic mechanism (in Luhmann's jargon, it is "structurally coupled"). "Agency" is the story we tell ourselves about this, but it is really an illusion (as Erich Hörl has powerfully discussed in his recent "The archaic illusion of communication")

By this mechanism, institutions conserve meaning. I wonder if they also conserve information, and Leydesdorff has done some very important work in applying Shannon's information theory to the academic discourse.

Ashby's insight into information systems becomes important: "Any system that categorises effectively throws-away information" he wrote in his diary. That seems perverse, because it means that our so-called information systems actually discard information! But they do.

For Luhmann, discarding information means that the probability that communications will be successful (i.e. serve the mechanism of autopoiesis in the institution) will be reduced. As he pithily put it in his (best) book "Love as Passion": "All marriages are made in heaven, and fall apart in the motorcar". What he means is that when one person in a couple is driving, their lifeworld is completely different to their partner's. The context for meaningful communication is impaired by the mismatch in communicative activity which each is engaged in.

In our social media dominated world, where alternative lifeworlds metastasise at an alarming rate, the effect of technology in damaging the context for the effective conservation of meaning is quite obvious.

In the technocractic world of the modern university, where computer systems categorise students with so-called learning analytics, it is important to remember Ashby: with each categorisation, information is thrown away. With each categorisation, the probability that communications will be successful is diminished as the sphere of permissible speech acts becomes narrower. Instead of talking about the important things that matter most deeply, conversations become strategic, seeking to push the right buttons which are reinforced by the institutional systems: not only the bureaucratic systems of the university, but the discourse system of the publishers, and the self-promotion system of social media. This is the real problem with data.

The problem seems quite clear: Our institutions are haemorrhaging information. It is as if the introduction of information systems was like putting a hole in the hull of the institutional "ship".

Stafford Beer knew this problem. It is basically what happens when the coordination and control part of his "viable system model" (what he called "System 3") takes over, at the expense of the more reflective and exploratory curious function that probes the environment and examines potential threats and opportunities (what he called "System 4"). In companies, this is the R&D department. It is notable that universities don't have R&D departments! Increasingly, R&D is replaced by "analytics" - the system 4 function is absorbed into system 3 - where it doesn't belong.

But let's think more about the technology. System 3 tools categorise stuff - they have to - it's part of what system 3 has to do. This involves selecting the "right" information and discarding the rest. It is an information-oriented activity. However, the opposite of information is "redundancy" - pattern, repetition, saying the same thing in many ways... in education, this is teaching!

Machine learning is also predominantly a redundancy-based operation. Machine learning depends on multiple descriptions of the same thing from which it learns to predict data that it hasn't seen before. I'm asking myself whether this redundancy-oriented operation is actually a technological corrective. After all, one of the things that the curious and exploratory function of system 4 has to do is to explore patterns in the environment, and invent new interventions based on what it "knows". Machine learning can help with this, I think.

But only "help". Higher level coordination functions such as system 4 require human intelligence. But human intelligence needs support in being stimulated to have new kinds of conversations within increasingly complex environments. Machine learning can be incredibly and surprisingly creative and stimulating. It can create new contexts for conversations between human beings, and find new ways of coordinating activities which our bureaucratic systems cannot.

My hunch is that the artists need to get on to this. The new institutional system 4, enhanced by machine learning, is the artist's workshop, engaging managers and workers of an organisation into ongoing creative conversation about what matters. When I think about this more deeply, I find that the future is not at all as bleak as some make out.

Tuesday, 5 November 2019

Non-Linear Dynamics, Machine Learning and Physics meets education

In my recent talk about machine learning (in which I've been particularly focussing on convolutional neural networks because they present such a compelling case for how the technology has improved), I explored the recursive functions which can be used to classify data such as k-means. The similarity between non-linear dynamics of agent-based modelling and the recursive loss functions of convolutional neural network training are striking. It is hard for people new to machine learning to understand that we know very little of what is going on inside. The best demonstration of why we know so little comes from demonstrating the non-linear dynamic emergent behaviour in an agent-based model. Are they actually the same thing in different guises? If so, then we have a way of thinking about their differences.

The obvious difference is time. A non-linear agent-based model's behaviour emerges over time. Some algorithms will settle on fixed points (if k-means didn't do this it would be useless), while other models will continue to feed their outputs into their inputs endlessly producing streams of emergent behaviour. The convolutional process appears to settle on fixed points, but in fact it rarely fully "settles" - one can run the python "model.fit()" function for ever, and no completely stable version emerges, although stability is established within a small fluctuating range.

I discussed this fluctuation with Belgian mathematician Daniel Dubois yesterday. Daniel's work is on anticipatory systems, and he built a mathematical representation of the dynamics that were originally introduced by biologist Robert Rosen. Anticipation, in the work of Dubois, results from fractal structures. In a sense, this is obvious: to see the future, the world needs to be structured in a way in which patterns established in the past can be seen to relate to the future. If machine learning systems are anticipatory (and they appear to be able to predict categories of data they haven't seen before), then they too will contain a fractal structure.

Now a fractal is produced through a recursive non-linear process which results in fixed points. This all seems to be about the same thing. So the next question (one which I was asking both Daniel Dubois, and Loet Leydesdorff who I saw at the weekend) is how deep does this go? For Loet, the fractal structures are in communication systems (Luhmann's social systems), and (importantly) they can be analysed using Shannon's information theory. Daniel (on whose work Loet has constructed his system), agrees. But when we met, he was more interested to talk about his work in physics on the Dirac equation and what he believes to be a deeper significance of Shannon. I don't fully understand this yet, but we both agreed that if there is a deeper significance to Shannon, then it was a complete accident because Shannon only half-understood what he was doing... Half-understanding things can be way forwards!

Daniel's work on Dirac mirrors that of both Peter Rowlands in Liverpool and Lou Kauffman in Chicago (and now Novosibirsk). They all know each other very well. They all think that the physical world is basically "nothing". They all agree on the language of "nilpotents" (things multiplying to zero) and quaternions (complex numbers which produce a rotational geometry) as the fundamental building blocks of nature. There is an extraordinary intellectual confluence emerging here which unites fundamental physics with technology and consciousness. Who could not find that exciting?? It must have significance for education!

What's it all about? The clue is probably in Shannon: information. And I think it is not so much the information that is involved in learning processes (which has always been the focus of cognitivism). It is the way information is preserved in institutions - from the very small institutions of friendship and family, to larger ones like universities and countries.

Our technologies are technologies of categorisation and they throw away information. Since the computer revolution, holes have appeared in our social institutions which have destabilised them. The anticipatory function, which is essential to all living things, was replaced with a categorising function. The way we use machine learning also tends to categorise: this would make things worse.  But if it is an anticipatory system, it can do other things - it can provide a stimulus for thought and conversation, and in the process put information back into the system.

That is the hope. That is why we need to understand what this stuff does. And that is why, through understanding what our technology does, we might understand not only what we do, but what our institutions need to do to maintain their viability.

Education is not really about schools and universities. Those are examples of institutions which are now becoming unviable. Neither, I think, is it really about "learning" as such (as a psychological process - which ultimately is uninspectable). Education is about "institutions" in the broadest sense: families, friendships, coffee bars, businesses, hospitals... in fact anywhere which maintains information. To understand education is to understand how the processes which maintain information really work, how they can be broken with technologies, and how they can be improved with a different approach to technology. 

Tuesday, 29 October 2019

About Aboutness and Relations: Thoughts on #TheDigitalCondition

As part of the Cambridge Culture, Politics and Global Justice group on the Digital Condition, I made a video response which sought to bring a cybernetic perspective to Margaret Archer's views on the "Practical domain" as pivotal in the relations between nature and the social. I remember challenging Archer on this many years ago when she gave a talk in London about her work on reflexivity and I suggested that Maturana and Varela's concept of "structural coupling" provided a clearer explanation of what she was trying to articulate in terms of the relations between people, practices and things. She brushed the point aside at the time, although more recently I heard her talk more approvingly of autopoietic theory, so I'd be interested to know what she thinks now. This is my video:



One of the things about making a video like this is that it is a very different kind of thing from  Archer's paper that we were all reading. Because it is more conversational, it expresses a certain degree of uncertainty about what it attempts to say. Not just in the messy diagrams, but in the pauses as I try to find the words for what I want to say. Also worth mentioning that having drawn the diagram, making the video was very quick. Why don't we do this more often? My suspicion is that as academics we are rather reluctant to reveal our uncertainty about things. Academic papers full of sophisticated verbiage are safer spaces to hide uncertainty. Personally, I think we should be doing the opposite to hiding uncertainty - and we have the technology to do it.

Anyway, this has elicited some defence of Archer - particularly in arguing that my critique is a misrepresentation of her argument. Well... I'm not sure.

In her paper she begins by focusing on "aboutness" and the relationship between consciousness and reality:
"Deprived of this reference to, or regulation by, reality, then self-referentiality immediately sets in – consciousness becomes to be conscious of our own ideas (generic idealism), experience equates knowledge with the experienced (pragmatism and empiricism), and language becomes the internal relationship between linguistic signs (textualism). Instead, consciousness is always to be conscious of something"
So "consciousness" is a thing which refers to another thing, "reality". So here are two distinctions. They are, of course, unstable and uncertain. What is reality? Well, what isn't reality?? What is consciousness? Well, what isn't consciousness?? (are rocks conscious, for example?) And if whatever is consciousness must refer to whatever is reality in order to be conscious, then what not-consciousness? Does that refer to anything? Lying behind all this is an implicit "facticity" behind the concepts of "consciousness", "reality" and "reference". Imposing the facticity effectively removes the uncertainty.

Archer says "consciousness has to be conscious of something", retreating from self-referentiality. But what if consciousness is self-referential? What does that do? It does two things:

  1. it creates a boundary, since self-reference is a circle.
  2. it creates uncertainty since whatever is contained in the boundary lies in distinction to what is outside it, where the nature of that distinction is unclear. Additionally, the totality of what is contained within the boundary cannot be accounted for within logic of the boundary (Gödel)

As I explain in my video, this then unfolds a topology.

So then what is reference? It must be about the way in which distinctions maintain themselves within the self-referential processes of consciousness. As I explain, this process entails transduction processes which operate both vertically (within the distinction) and horizontally (between a distinction and its environment).

There's a very practical example of this from biology. One of the central questions about DNA is "How does a molecule come to be about another molecule?" (thanks to Terry Deacon for that!). This is a profound question which throws into doubt what is known as the "central dogma" of biology, which places DNA at the centre of life. It really can't be right.

What is more likely is that there are processes of bio-chemical self-reference initially involving lipid membranes maintaining their internal organisation and boundaries in the context of an ambiguous environment. DNA can then be seen as an epiphenomenon of the evolution of this communicative process. In other words, the aboutness of DNA is our distinction concerning the emergent epiphenomena of self-reference.

It's the same with reference more generally. Once we see consciousness as self-reference, then the categories that we invent about "nature" or "the social" can be seen as epiphenomena of the self-referential process. This makes explaining this stuff a lot simpler, in my view. (And of course, explanation is another epiphenomenon of self-reference). It also helps to explain the ways in which we bring technology to bear on processes which help us to maintain our distinctions.

Wednesday, 23 October 2019

Ancestrality and Consciousness

Over the past year, I've been increasingly convinced of the correctness of the evolutionary biological theory of John Torday concerning the connection between consciousness, cellular evolution and "big" evolutionary history (from the deep origins of space and time). Of course, it's hugely ambitious - but we should be hugely ambitious, shouldn't we?

John's work in physiology and medicine (primarily focused on lung physiology and asthma) has presented a number of empirical phenomena that point towards a biological theory that includes evolutionary history, where consciousness is part of a process that explains how cells evolve from lipid bi-layers to sophisticated inter-cellular communication systems. It also addresses what he, and many other biologists see, as a scientific problem within their discipline - that it is not explanatory in the way that physics or chemistry describe causal mechanisms, but descriptive. In our extensive conversations, I have noted that education suffers the same problem: there is no mechanistic explanation for educational phenomena, only description. Since education is also a manifestation of consciousness, I am concerned to make the connection between these different lines of inquiry (biology, physics and evolution).

The central question in evolution is the relationship between diachronic (time-based) process and synchronic structures. What time-based process makes cells absorb parts of their environment (bacteria for example) as mitochondria? What time-based process introduces cholesterol as the critical ingredient to animate life? What time-based process governs the expression of proteins and their transformations in the cell signal-transduction pathways, which despite their complexity, maintain the coherence of the organism?

Time itself introduces further questions. When we look at evolutionary history - maybe at the red-shift of the expanding universe - what are we looking at exactly? "Once upon a time, there was absolutely nothing, and then there was this enormous explosion..." really?

I've been re-reading Quentin Meillassoux's "After Finitude". I have some misgivings about it, but this must surely be one of the greatest philosophical works of the last 20 years. The question of time is one of his central questions, and he calls it "ancestrality". The question is about the nature of reality, and particularly the reality of things like fossils, or electromagnetic radiation from outer-space. We either assume that the world is made by our consciousness, or we assume that the world exists in a pre-existing domain that exists independently of human consciousness and agency (what Bhaskar calls the "intransitive domain"). Meillassoux pursues the Platonist position which denies both (in line with Alain Badiou) arguing that objects are mathematically real to us - logic, in other words, is prior to materiality. At the heart of Meillassoux's (and Badiou's) argument is the contingency of nature. He asks, "Given this contingency, how do things appear stable?"

Pursuing this, the ancestrality of the universe - the big bang, evolutionary history - is (Meillassoux would claim) "logically" real. But this puts the emphasis on the synchronic reality of things - their logical structure out of time - and it assumes that the consciousness that conceives of this logic is similarly structured. Indeed, I'm not convinced that one of Meillassoux's central points - that the Western philosophical position is one of "correlation" between ideas and reality is escaped in his own position. But the care with which he lays out his arguments is nevertheless highly valuable, and his emphasis on contingency seems right to me (but maybe not! - how can anyone say an ontology of contingency is "right"?)

Torday's situating of time in the material and biological evolution of consciousness means that this "logic" has to become a "topo-logic": space and time - the diachronic dimension - are not separable from the "logic" of synchronic structure. What do we get when we have a "topo-logic"? We get contingency, process, uncertainty and the driving necessity for coherence. In essence, we get life.

Somehow, we have to grapple with topology. For a long time, I struggled with the concept of time within cybernetics. After all, you have to have time to have a mechanism, but where did "time" come from? There must be something prior to "mechanism". It turns out that when we think through one of the other key distinctions of cybernetics - difference - we find the answer. A difference results from a distinction. A distinction is a boundary which marks what is inside from what is outside. But distinctions are essentially unstable: whatever mark is made generates a question. It's the same question that Gödel addressed: the distinction demarcates a "formal system", but there are propositions expressible within the formal system which cannot be proved within that system. Uncertainty is inevitable - how is it managed?

Something must be invented in order to mop-up the uncertainty. Time is a powerful invention. By creating past, present and future, ambiguities can be situated in a way where contradictions can be expressed "now it is x" and "now it is not-x". The implications of this are that the topology becomes richer as the distinctions about time must also be negotiated. Part of the richness of this topology may also be the creation of deep symmetries in time and space, including concepts such as "nothingness" or nilpotency, and "dimensionality" - in essence, as my colleague Peter Rowlands would argue, the foundational principles of the material world. The invention of time entails a double management process, where part of it must coordinate with an environment which is also the cause of uncertainty. There are many distinctions in the universe, each creating time, space and matter, and each constraining other distinctions.

So is the "big bang" story a manifestation of a topology? Does the topology pre-exist the consciousness which conceives of it? (what does "pre-" mean in that sentence?) If this is so, what is "evolution"? I feel myself skirting foundationalism while denying the possibility of any "foundation"... and then seeing that process as a foundation... then denying it... then seeing it as a foundation... and so on.

"In the beginning was the word" says St. John's Gospel. That's a distinction - it unfolds a topology. Theologians like Arthur Peacock imagined that "logos" might also mean "information". If there is information, then there is topology, and then the "beginning" is "the word" - the distinction.  And beginnings are everywhere, not least the beginnings created by the distinctions of consciousness. But consciousness's beginnings have their roots in the beginnings of matter - in the "word".

We're very close to Torday's essential point: cells, from which consciousness emerges, are stardust which trace their evolutionary history to the beginning. In the topology of maintaining distinctions, new distinctions must be made as the ambiguity of the environment is dealt with. Indeed, the difference between atoms and organisms may be that atoms in maintaining their distinction, must find a way of organising themselves such that new distinctions may be made. The environment within which atoms organise is the essential driver for new forms of organisation. That way of organising is what life is: a search for new ways of making distinctions which manage the uncertainty generated by those distinctions. It is this, I suspect, which is the mechanism of evolution. It's only about history in the sense that our unstable distinctions require us to invent history to maintain our distinctions about ourselves and our environment.

What we call "homestasis" is the cell's drive for coherence in its distinction-making. What we call "information" or "negentropy" is the cell's interface with its environment. What we call "chemiosmosis" is the disturbance to the cell's equilbrium by forces in its environment and its gaining of energy.

Thought itself is the universe's way of making new distinctions. Since the universe is imprinted in the biology of consciousness, the symmetries of physics, biology and consciousness will contrive to form coherences which enfold both synchronic and diachronic dimensions. This may be why the crazily complex protein dance hangs together - because of the deep coherence between diachronic and synchronic dimensions.

What then of science itself? Of empiricism? In a world produced by thought, something happens within the time we invent to establish coherence between thought and the world. Thought looks closely at what it has made, it discards certain aspects of what it sees, it executes control on what is left, it observes what happens - not just one mind, but many - and then collectively it thinks more deeply. A new level of coherence is arrived at and the topology unfolds once more. 

Sunday, 6 October 2019

Creatively defacing my copy of Simon Critchley's "Tragedy, the Greeks and Us"

I've been defacing my copy of Simon Critchley's "Tragedy, the Greeks and Us". For me, this vandalism is a sign that something has got me thinking. It's not just Critchley. I went back to Jane Harrisson's "Ancient Art and Ritual" the other day, partly in response to my recent experiences in Vladivostok and a central question concerning the structure of drama and the structure of education. Basically: is education drama? Should it be? and, Is our experience online drama? Critchley's not dismissive of Harrison and the Cambridge ritualists - which I find encouraging - and I like his suggestion that art may not be so much "ritual" as "meta-ritual". 

It's funny how things revolve. I was introduced to Harrison by Ian Kemp at Manchester university as a student, who was also a passionate expert on Berlioz. Yesterday evening I took my daughter to hear a performance of Berlioz's Romeo et Juliette, which is Berlioz's brilliant and beautiful refashioning of Shakespeare into the form of a symphony via Greek drama: it has explicit sections of chorus, prologue, sacrifice, feast, etc. Beethoven meets the Greeks!

I'm very impressed with Critchley - and I very much get his vibe at the moment - that tragedy and the ambiguity of dramatic structure was overlooked in favour of philosophy (Plato particularly), and that we are now in a mess because of it. I agree. If we replace "tradegy" with "the drama of learning" or "the dialectic of self-discovery" then I think there are some important lessons for education. Critchley makes the point that our modern lives are determined by endless categorisation, and the resulting incoherence of this drives us back to Facebook and social media:
"We look, but we see nothing. Someone speaks to us, but we hear nothing. And we carry on in our endlessly narcissistic self-justification, adding Facebook updates and posting on Instagram. Tragedy is about many things, but it is centrally concerned with the conditions for actually seeing and actually hearing"
That's what I was missing in "The Twittering Machine". 

But he has an axe to grind about philosophy and Plato - and particularly with his contemporary philosophers, most notably Alain Badiou. Since Badiou also has a deep interest in the arts (and opera particularly) this is interesting, and I think Critchley is seeing a dichotomy where there isn't one. And that is where my doodling starts...

The essence of this goes back to the relationship between the synchronic, categorical frame of rationality and experience which demarcates times, and the diachronic, ambiguous frame which sees time as a continuous process. Critchley doesn't seem to see that the two are compatible. But I think they are in a fundamental way. 

The issue concerns what a distinction is, and the relationship of a distinction to time. We imaging that distinctions are made in time, and that time pre-exists any distinction. But it is possible that a distinction - the drawing of a boundary - entails the creation of time. So this was my first doodle:


The distinction on the right is simply a self-referential process. It embraces something within it, but it occurs within a context which cannot be known (in this case, a "universe" and an "earth" and an "asteroid"...) All of those things are distinctions too, and they are all subject to the same process as I will describe now. The essential point is that all distinctions are unstable.

When we think of distinctions as unstable, think of a distinction about education. There are in fact no stable distinctions about education. Everything throws us into conversation. In fact the conversation started long before anyone thought of education. Indeed, it may have started with the most basic distinctions of matter.

If I was to draw this instability, it is a dark shadow emerging within the distinction. These are the unstable forces which will break apart the distinction unless they are absorbed somehow. So we need something to absorb the uncertainty. It cannot be inside the distinction - it must be outside. So we are immediately faced with a duality - two sides of the Mobius strip.

But more than that, absorbing uncertainty (or ambiguity if you want) is a battle on two fronts. The internal uncertainty within the distinction is one thing, but it must be balanced with what might be known about the environment within which the distinction emerges and maintains itself.

Let's call this "uncertainty mop" a "metasystem".  And here it is. Note the shadow in the distinction and the shadow in the environment. Part of me wants to draw this like a Francis Bacon "screaming pope".


A war on two fronts is hard - the metasystem better get its act together! Part of it must deal with the outside and part of it must deal with the inside - and they must talk to each other.

The interesting and critical thing here is that in order to make sense of this balancing act, the metasystem has to create new distinctions: Past, Present and Future. We might call these "imaginary" but they are entirely necessary, because without them, there is no hope of any kind of coherent stability in our distinction. But what have we done? Our distinction has made time!


By inventing time, we have invented the realm of the "diachronic". This is the realm of drama and music. Whatever time is - and how can we know? - it expands our domain of distinction-making, and helps us to see the connection between past, present and future. In the language of "anticipatory systems" of Daniel Dubois and Robert Rosen, this is the difference between recursion (the future modelled on the past), incursion (the future modelled on the future), and hyperincursion (selection among many possible models of the future).




I ought to say that all of this happens all at once. A distinction is like a bomb going off - or even a "big bang". But these bombs are going off all the time - or rather, they are going off all the time that they themselves make. A distinction entails time, which entails dialectic (and conversation). But it also entails hope for a coherence and stability of a distinction within what it now sees as a changing world.

I think the best picture of coherence is the relationship between a fractal as a kind of map of the environment and the unfolding patterns of action within that environment. It's a fractal because only a fractal can contain seed of the future based on its past. Life goes on in the effort to find a coherent structure. When it does, we die.
Mostly, the search for coherence leads to new distinctions, and so the process goes on in a circle. This, I think, is what education is.

The structure of tragedy unfolds this circle in front of us for us to see. It is a circle of nature - of logic. It is the logic of every atom, cell, fermion, quark, whatever... in the universe. It is the logical structure of a distinction. 

When Critchley says of tragedy that it is about "actually seeing and actually hearing" he is spot-on. But I think his anti-Platonist stance is a reaction to where we are now. "Actually seeing and actually hearing" has been replaced with the processing of data. The part of the metasystem which does that is the lower-part, mopping up the internal uncertainty, but not really thinking about the environment. The diachronic bit - the time-making bit - has been crowded-out by our computer powered categorisation functions. If we continue like this, hope will be extinguished, because hope also sits in the upper bit.

Biological systems do not suffer this imbalance. We have had it forced on us by our rationality. That is our tragedy. Yet we are not as rational as we think - and our rationality is a biological epiphenomenon. It feels as if our technology is out of balance: a dialectical imbalance that presents us with a challenge to be overcome for the future.  But this may be as much of a question as to how we organise our institutions as it is about what kind of technology we produce in the future. 

We have the option to organise our education differently, and learn something from the past. We also have the option to use our technologies differently and use them to reinforce the ambiguities of distinctions, rather than the information-discarding processes of categorisation. 

Wednesday, 2 October 2019

Design for an Institution

To paraphrase Ashby's "Design for a brain", it might be asked "How does an institution produce adaptive behaviour?" But of course, institutions often don't produce adaptive behaviour, or their adaptations lead them to adopt patterns of behavior which are more rigid - which is the harbinger of death for an organism, but institutions seem to be large enough to withstand environmental challenges that they continue to survive even without apparently adapting.

One way of answering this problem is to argue that institutions, in whatever form they come, and in whatever state of adaptability, conserve information. An institution might lose a quantity of information in response to an environmental threat, but maintain some stable behaviour despite this. It's internal processes maintain that new set of information. This is, I think, like what happens when institutions are challenged by a complex environment and become more conservative. They discard a lot of information and replace it with rigid categories, often upheld by computer technology and metrics. But the computer-coordinated information-preservation function with a limited information set is surprisingly resilient. However, for human beings existing within this kind of institution, life can be miserable. This is because human beings are capable of far richer information processing than the institution allows - effectively they are suppressed. There may be distinct phases of information loss and preservation.

But if institutions are information-preserving entities, then rigid low-information preserving entities will not be able to compete with richer information-preserving entities. If information preservation is the criteria for "institution-ness", then new ways of preserving information with technology may well be possible which might challenge traditional institutional models. So what is a basic "design for an institution"?

An institution, like a brain, must be a collection of components which communicate - or converse - with one another. In the process of conversation, essential distinctions about the institution are made: what is it, what is it for, what functions must it perform, and so on. Each of these distinctions is essentially uncertain: conversation is necessary to uphold each distinction. Within the conversations there are details about different interpretations of these distinctions. Not all of these differences can be maintained: some must be attenuated-out. So information is lost in the goal of seeking generalisable patterns of practice and understanding to coordinate the whole.

Having said this, the generalisations produced may well be an inadequate representation. So what must be done is that whatever generalisations are produced are used to generate multiple versions of the world as it is understood by these generalisations. The multiplicity of this generated reality and the multiplicity of "actual" reality mus be compared continuously, and the question asked "in what ways are we wrong?" This generation of multiple descriptions is a niche-making function that can generate new information about the world. It is like a spider spinning a web to create a home, but also to detect what is in the environment.

An institution will connect the generalisations it makes with the new information produced through its multiple expressions of its understanding. In traditional institutions, both functions were performed by humans: an "operations" team that identified what needed to be done and did it; and a "research and development" team which looked at the future and speculated on new developments. Technology has shifted the balance between operations and strategy, where research and development is now seen as "operational" in the sense that it has become "data driven". This collapse of distinction-making is dangerous.

But lets say, in a technology-enhanced institution (such as all of ours are now), a clear division could be established between the operational, synergistic parts of the organisation, and the strategic, future-looking parts. We need two kinds of machines. One, the purpose-driven analytical engine that the modern computer is, in order to maintain operations and synergy. The other, a "maverick machine" as Gordon Pask put it, which uses the information it has at its disposal to produce a rich variety of artefacts from paintings and music to usual correlations of data. The maverick machine's purpose (if it can be said to have one) is to stimulate thought and shake it out of the rigid confines of the analytical engine. It presents people with orders of things which are unfamiliar to them, and challenges them to correct it, or embrace a new perspective. By stimulating thought, the maverick machine generates new information to counteract the information that is lost through the analytical engine. The maverick machine is an information-preserving machine  because it maintains a living record of human order and distinction-making within the institution.

It is through the stimulation of thought that changes might be made to the analytical engine and new strategic priorities defined, or new orders identified. In this way, an institution must be a collaboration between humans and machines. To think of institutions as essentially human, with technological "support" is a mistake and will not work.

What is particularly interesting about the maverick machine is that it is a creating entity. Not only does that mean that the viable institution maintains its information. It also means that the viable institution is putting stuff out into the environment in the form of new things. This "generous" behaviour will result in patterns of engagement with the environment which will help it to survive. People will support the institution because not only does its information-preserving processes help itself, but helps other things and people in the environment too.

What does this actually look like? Well, imagine two friends who have similar intellectual interests. They meet every now and then and discuss what they are reading and are interested in. But they don't just discuss it, they video their discussions. They process the video to extract text and images. They use machine learning to mine the text and explore new resources, which software is then able to produce new representations of (a maverick machine). A weekly meeting is generative of a rich range of different kinds of things. Others see these things and think "that's cool - how do I get involved?" Using the same techniques, others are able to do similar things, where the software is able to create synergies between them. Slowly an information-preserving "institution" of two friends becomes something bigger.

This is not Facebook: that is an institution which loses vast amounts of information. It is more akin to a university - an institution for preserving information and creating the conditions for conversation.  

Monday, 30 September 2019

Technology and the Institution of Education

There's a lot of stuff about technology in education on the internet at the moment. A lot of it is increasingly paranoid: worries about the "platform" university, surveillance, brain scanning, boredom scanning, omniscient AIs, idiot AIs, and big corporations making huge sums of money out of the hopes and dreams of our kids.  Whitehead noted once that if one wants to see where new ideas are going to arise, one has to see what people are not talking about. So, most likely the future is going to be "none of the above". But what are we not talking about?

The Golem-like AI is, and always has been, a chimera. What is real in all this stuff? Well, follow the money. Educational institutions are enormous financial concerns, boosted by outrageous student fees, burgeoning student numbers, increasingly ruthless managerial policies of employment, and an increasing disregard for the pursuit of truth in favour of the pursuit of marketing-oriented "advances" and "high ranking" publication - bring on the Graphene! (condoms, lightbulbs, and water purification here we come. perhaps.) Of course, tech companies are massive financial concerns too, but while we are all on Facebook and Twitter, we are not taking out thousands of pound loans to feed our Facebook habit. Naturally, Facebook would like to change that. But it seems a long-shot.

So we come back to the institution of education. Why has it become such a dreadful thing? When I think about my own time in the music department at Manchester University in the late 80s, I think of how my best professors would probably not be able to get a job in the metrics-obsessed University now. This is a disaster.

Ross Ashby (another genius who would have struggled) noted that any system that distinguished categories effectively "throws away information". The profundity of that observation is worth reflecting on. All our educational IT systems, all our bibliometric systems, NSS, REF, TEF, etc are category-making systems. They all throw away information. The result of the impact of technology - Information technology no less - in education has been the loss of information by its institutions.

What happens to institutions when they lose information? They lose the capacity to adapt their operations in an increasingly complex environment. As a result, they become more rigid and conservative. Information systems create problems that information systems can solve, each new wave of problems loses more information than the previous wave. We are left with a  potentially irrelevant (although still popular) operation which has little bearing on the real world. This is where we are, I think.

Let's be a bit clearer about institutions: institutions which lose information become unviable. So, a viable institution is an entity which conserves information. Traditionally - before technology - this was done by balancing the institution's operational (information losing) function with a reflexive (information gaining) function that would probe the environment, where academics had space for thinking about how the world was changing and making informed interventions both in the world and in the institution. When technology entered the institution, the operational function - which was always a "categorising" function - was amplified, and to many excited by the apparent possibilities of new techno-coordinating powers, the loss of information was welcomed, while at the same time, the reflexive function was dismissed as a waste or irrelevant. Basically, everything became "operations", and thought went out of the window.

Many AI enthusiasts see AI as a further step towards the information-losing side of things, and welcome it. AI can lose information better than anything - basically, a technology for soaking up a large amount of information redundancy for the sake of producing a single "answer" which saves human beings the labour of having to talk to each other and work out the nuances of stuff.

But in the end, this will not work.

But AI, or rather deep learning, is a different kind of technology. In working with redundancy rather than information, it is something of a counterbalance to information systems. Redundancy is the opposite of information. Where information systems amplified the "categorising" operational processes, might deep learning technology amplify the reflexive processes? I am particularly interested in this question, because it presents what might be a "feasible utopia" for technologically-enhanced institutions of education in the future. Or rather, it presents the possibility of using technology to conserve, not destroy, information.

The key to being able to do this is to understand how deep learning might work alongside human judgement, and particularly the ordering of human judgement. If deep learning can operate to preserve human order, coordinate effective human-based correction of machine error, whilst supporting human judgement-making, then a virtuous circle might be possible. This, it seems to me, is something worth aiming for in technologically embedded education.

Conserving information is the heart of any viable institution. States and churches have survived precisely because their operations serve to preserve their information content, although like everything, they have been challenged by technology in recent years.

In Stafford Beer's archive, there is a diagram he drew of the health system in Toronto. At the centre of the diagram is a circle representing a "population of healthy people". This is a system for preserving health, not treating illness. And more importantly it is a system for preserving information about health.


We need the same for education. In the centre of a similar diagram for education, perhaps there should be a box representing a population of wise people: everyone from children to spiritual leaders. What is the system to preserve this wisdom in society? It is not our current "education" system that seeks to identify the deficit of ignorance and fill it with lectures and certificates. Those things make wise people go mad! It is instead a set of social coordination functions which together preserve information about wisdom, and create the conditions for its propagation from one generation to the next. We don't have this because of the information loss in education. Can we use technology to correct the loss of information and turn it into conservation? I think we might be able to do this. We should try.

My concern with the paranoia about technology in education at the moment is that it is entirely framed around the perspective of the traditional institution in its current information-losing form. It is effectively realising that the path of information-loss leads to the abyss. It does indeed. But it is not technology that takes us there. It is a particular approach to technology which turns its hand to categorisation and information loss which takes us there. Other technologies are possible. Another technological world is possible. Better education, which preserves information and maintains a context for the preservation of wisdom between the generations is possible. But only (now) with technology. 

Tuesday, 24 September 2019

Time, Ritual and Education

The Global Scientific Dialogue course at the Far Eastern Federal University, which I co-designed with Russian colleagues last year, is running for a second year. I was happy with how it went last year, but this year it seems better. I must admit that on returning to this transdisciplinary mix of stuff (art, science, technology, intersubjectivity, machine learning, etc), I had a few worries about whether I was deceiving myself that any of this was any good – particularly as this year, we are running it with nearly 500 students. But after getting back into it, and particularly after talking to the teachers, I was reassured that there was something really good in it, which was of special benefit to teachers and students across the management and economics school in the university.

Although the course is about trying to provide a broader perspective on the rapid changes in the world of work for our students (particularly, this year, with a focus on machine learning), I think this is really as much a course for teachers: it demands and gets great team teaching. Last year we recruited and trained 30 teachers from the school, and 20 helped us run the course. This year we recruited and trained another 17. But it was so much easier because last years’ teachers have become experts: from being a very small team trying to encourage innovative teaching practice (basically me and a couple of Russian colleagues), it has been transformed into a movement of more than 20 teachers all pulling in the same direction. Their internal communication has been conducted through WhatsApp, and this year, the level of cooperation and coordination has been superb. It’s really wonderful. Eventually, they may not need me any more – but that’s as it should be!

Technologically, it’s very simple. There is video to keep things coordinated so many teachers can conduct similar activities in small groups together, there is comparative judgement to keep the students thinking and submitting their thoughts, and patchwork text to provide flexibility in assessment. We tried as hard as we could to get away from rigid learning outcomes. We ended up with a compromise.

Ok. So it really works (although much could be refined). Why?

I gave a presentation to the senior academic board of the faculty last week. I explained that I see the course as a cybernetic intervention, inspired by Beer’s work on syntegration. But only inspired. Really, I think the interventions of the course all contribute to an uncertain environment for teachers and students (syntegration does this too). The uncertainty means that they cannot rely on their pre-existing categories for dealing with the world (existing within what Beer calls the “meta-system”), and must find ways of reconfiguring their meta-system, expressing their uncertainty, which they do through dialogue. Importantly, this helps to level the positioning between teachers and students.

I’m fairly happy with that as an explanation: the evidence fits the model. But I think there’s something more. I’m wondering if the course’s structure over its two weeks is also important.

The structure is highly varied. It begins with a “big lecture” from me. I was never that comfortable with this but its an administrative requirement, and the room we do it in is enormous and echoey. So I start by getting them all to sing, and I introduce the idea of multiple description through examining the sound frequencies in a single sound (“A single sound is made of many frequencies. A single concept is produced by many strands of dialogue, etc…”). It’s great having a spectrum analyser providing real-time feedback and intellectual challenge!

Thinking about it, the singing is a “chorus” - does all this have the structure of ancient Greek drama?

I try not to talk for too long before getting them to turn their chairs round and play a game. We play Mary Flannagan’s "Grow-a-game", asking the students to invent a new game that addresses a global challenge (inequality, homelessness, global warming, etc) by changing the rules of a game they already know.

There is much argument and debate in the groups. A kind of “Agon” (bear with me….)
We get the students to make a short video of their game, which we play to everyone. This is a presentation of ideas and themes, maybe a “Parados?

More talk follows (chorus), followed by games (agon 2).

Then students go to separate groups and talk about different topics. These topics too have a similar structure, except the “chorus” is usually a video that sets the scene  (could be a “prologue”). Then there is more activity (agon 3), and a presentation of their ideas ("stasimon"?)

In the middle of the course, we have a “feast”: a gathering of experts where all 500 students are free to wander around and talk to interesting people. I told the students to think of it as a “party”. It wasn’t quite a party, but had a great atmosphere for everyone.





At the end of the course, the students parade their work. We end with a final procession ("exodus"?).

Maybe I’m being overly grand, but the thing has a structure, and I can’t help feeling that the structure has a deeper significance which may relate to ancient drama and ritual - although things may be in an unconventional order in comparison to Euripides.

What was the point of ancient drama and ritual? It must have had a function in producing coherence of experience among the group of spectators. That is exactly what we are trying to do with Global Scientific Dialogue. Why is so much formal learning incoherent? Because it doesn’t have any clear structure: it’s just one thing after another. There’s no dramatic thrust. This may explain why some talk about the "accelerated academy" (like this: http://accelerated.academy/). I don't buy it - "acceleration" is the feeling one gets when things start to run out of control. The real problem is that things just don't make sense. Our traditional ways of operating in education are out of control.

The connection must be made between the message that is given, how it is given, the structure of proceedings, and the role of time. So what is the connection?

We can think of a message as a distinction. When we draw a distinction we immediately create uncertainty: what is inside and what is outside the distinction? The internal uncertainty must be managed and balanced with external uncertainty. In order to manage the external uncertainty, the invention of time is necessary. It's only by creating past, present and future that the essential contingency of a distinction can be maintained.

With the creation of past, present and future, the relationship between diachronic structure and synchronic structure becomes an important element in the coherence of the whole: exactly in the way that music operates. Could it be shown that the moments of the dramatic ritual are necessary to maintain coherence? Well, these elements like “agon”, “parados”, “feast”, “chorus” rotate around in different configurations. They are discrete moments produced by dialogue. Each is characterised by a different form of pattern or "redundancy". I suspect these are different ways of saying the same thing - or maybe different ways of saying "nothing". What is the switch from one “nothing” to the next “nothing”?

I'm thinking that this brings Peter Rowlands’s idea from physics of mass, space, time and charge all dancing around each other in a cyclic group is useful. I strongly suspect there is some deep contact between the nature of the universe and the structure of the moments of experience. We could go much deeper.

But whatever theoretical construct we might indulge in, Global Scientific Dialogue has presented a phenomenon which demands a better explanation than “everyone seems to like it!”.

Monday, 16 September 2019

Topology and Technology: A new way of thinking

I'm back in Russia for the Global Scientific Dialogue course. We had the first day today with 250 first year students. Another 250 second year students follow on Wednesday. They seemed to enjoy what we did with them. It began with getting them to sing. I used a sound spectrum analyzer, and discussed the multiplicity of frequencies which are produced with any single note that we might sing. With the spectrum analyzer, it is possible to almost "paint" with sound: which in itself is another instance of multiple description. A very unusual way to begin course on management and economics!

The message is really about conversation, learning and systems thinking. Conversation too is characterised by multiple descriptions - or rather the "counterpoint" between multiple descriptions of things. This year, largely due to my work on diabetic retinopathy, the course is spending a lot of time looking at machine learning. Conversations with machines are going to become an important part of life and thanks to the massive advances in the standardisation of ML tools (particularly the Javascript version of tensorflow meaning you can do anything in the web), you don't have to look far to find really cool examples of conversational machine learning. I showed the Magic Sketchpad from Google fantastic Magenta project (a subset of their Tensorflow developments): https://magic-sketchpad.glitch.me/. This is clearly a conversation.

It feels like everything is converging. Over the summer we had two important conferences in Liverpool. One was on Topology - George Spencer-Brown's Laws of Form. The other was on physics (Alternative Natural Philosophy Association) - which ended up revolving around Peter Rowlands's work. The astonishing thing was that they were fundamentally about the same two key concepts: symmetry and nothing. At these conferences there were some experts on machine learning, and other experts on consciousness. They too were saying the same thing. Symmetry and nothing. And it is important to note that the enormous advances in deep learning are happening as a result of trial and error, and there is no clear theoretical account as to why they work. That they work this well ought to be an indication that there is indeed some fundamental similarity between the functioning of the machine and the functioning consciousness.

My work on diabetic retinopathy has basically been about putting these two together. Potentially, that is powerful for medical diagnostics. But it is much more important for our understanding of ourselves in the light of our understanding of machines. It means that for us to think about "whole systems" means that we must see our consciousness and the mechanical products of our consciousness (e.g. AI) as entwined. But the key is not in the technology. It is in the topology.

Any whole is unstable. The reasons why it is unstable can be thought of in many ways. We might say that a whole is never a whole because something exists outside it. Or we might say that a whole is the result of self-reference, which causes a kind of oscillation. Lou Kauffman, who came to both Liverpool conferences, draws it like this (from a recent paper):


Kauffman's point is that any distinction is self-reference, and any distinction creates time (a point also made by Niklas Luhmann). So you might look at the beginning of time as the interaction of self-referential processes:
But there's more. Because once you create time, you create conversation. Once the instability of a whole distinction is made, so that instability has to be stabilised with interactions with other instabilities. Today I used the idea of Trivial Machine, proposed by Heinz von Foerster. Von Foerster contrasted a trivial machine with a non-trivial machine. Education, he argues, turns non-trivial machines into trivial machines. But really we need to organise non-trivial machines into networks where each of them can coordinate their uncertainty.
I think this is an interesting alternative representation of Lou's swirling self-referential interactions. It is basically a model of conversation.

But this topology opens out further. Stafford Beer's viable system model begins with a distinction about the "system" and the "environment". But it unfolds a necessary topology which also suggests that conversation is fundamental. Every distinction (the "process language" box) has uncertainty. This necessitates something outside the system to deal with the uncertainty. If we assume that this thing outside is dealing with the uncertainty, then we have to assume that it must both address uncertainty within the system, and uncertainty outside it. Since it cannot know the outside world, it must perform a function of probing the outside world as a necessary function of absorbing uncertainty. Quickly we see the part of the system which "mops up" the uncertainty of the system develops its own structure, and must be in conversation with other similar systems...


What does this mean?

Beer's work is about organisation, and organisation is the principle challenge we will face as our technology throws out phenomena which will be completely new to us. It will confuse us. It is likely that the uncertainty it produces will, in the short run, cause our institutions to behave badly - becoming more conservative. We have obvious signs right now that this is the case.

But look at Beer's model. Look at the middle part of the upper box: "Anticipation". Whole distinctions make time, and create past and future. But to remain whole, they must anticipate. No living system cannot anticipate.

With the rapid development of computers over the last 80 years, we have had deterministic systems. They are not very good at anticipation, but they are good at synergy and coordination (the lower part of the upper box). But we've lacked anticipation - having to rely on our human senses which have been diminished by the dominance of deterministic technology.

I could be wrong on this. But our Deep Learning looks like it can anticipate. It's more than just a "new thing". It's a fundamental missing piece of a topological jigsaw puzzle.