Music Black-Boxes

Everyone knows what a music box is - when we turn the handle a known tune will start at some random point. We understand the mechanism inside. The charm is perhaps that despite understanding the mechanism, the music itself can make us feel different at different moments. The box might be predictable and mechanical, but we are not.

This has got me thinking about music itself and the somewhat mechanical way we are taught to think about it. We are told of the component parts of music: there is rhythm, harmony, melody, tonality, dynamics, etc. Music is such a difficult subject because all of these things are incredibly abstract: where is the difference between harmony and counterpoint, for example? Ian Kemp, my professor of music at Manchester would often say, "well, perhaps it's just a nice noise".

We only think we understand the components of the mechanism of the music box. Just as we only think we understand the components of music itself. In both cases, we lose sight of ourselves as the context within which the understanding takes place. Imagine the music box is broken down into independent boxes: one for the crank, one for the chimes, one for the sprockets, one for the barrel, etc. To make the music, these components have to 'communicate' in the right way. The glue that makes things work is us: and we are not predictable and mechanical in the way we imagine the components to be. Indeed, the components and their interconnections are a product of our unpredictableness.

To see how we invent components and their relations, I've found it useful to consider not the components as "things" (like melody, harmony, etc), but simply as 'black boxes': entities whose behaviour is obscure, and whose mechanisms cannot be inspected, but which define themselves through their interactions with us and with other black box entities. A music black box is an entity with behaviours which we can only speculate on through observation. As we speculate we may in fact create new kinds of black box: rhythm, harmony, melody, and so on. Doing so creates more things to distinguish any particular black box against.

What's interesting me is that all intellectual inquiry about things such as music creates multiple black boxes with various inter-connections. A black box establishes itself in its contradistinction to other black boxes. It is not in itself real; it is only real in its relations. We can do analysis or research on this by making explicit judgement about particular kinds of transactions which one boxes engages with another. So there might be rhythm transactions, and harmony transactions, and melody transactions and so on.

By determining a body of transactions and associating it with a particular black box we can speculate on the dynamics of the interaction between an interconnected chain of black boxes.  We are part of the chain, and the 'nice noise' emerges from the ensemble.

Where does this take us?

The important thing here is the relations between black boxes. Each box exercises constraint on the others, and each produces the conditions for growth in the others. There is mutuality of constraint in their relations and there is a kind of mutual catalysis. This may be like Bateson's distinction between 'symmetrical schismogenesis' and 'complementary schismogenesis'. There are empirical ways we can analyse this through information theory. Each component transforms and is constrained by a lifeworld of expectations and events which affects all components. The environment contains the realm of the 'possible' - and this realm changes are things progress. It is the equivalent of 'maximum entropy' in information theory. So the rhythm of 'Bolero' constrains other elements like harmony, melody, orchestration in a way which maintains maximum entropy; the Tristan chord shapes melody and tonality (particularly) in a way which increases it (this is why it was revolutionary).


Every distinction we trace through a pattern of growth and development takes the form of a black box defined in contradistinction to others on account of its 'transactions' with its neighbours and its environment. Studying the transactions between things: the mutuality of constraint, the autocatalysis can give some indication of the underlying conditions within which things grow - including us. Understanding how we grow can help us understand the conditions for creating new distinctions and new black boxes...

Stafford Beer's comments on a societal meta-language

As I've been contemplating the impact of Block Chain on education, and thinking about how new kinds of transactional systems might transform the way we think about our institutions, my attention has been drawn back to Stafford Beer's book "Platform for Change'. Beer attempted to articulate a cybernetic social transformation where he speculated on the move from old institutions to new, cybernetic, institutions. At the heart of a new world was the development of a "meta-language" which would describe the new society in terms of ecological dynamics. That brings new categories of thinking about society, work, institutions, etc. and it brings new kinds of metrics. Beer's motivation was driven by the existential fears of the 1970s. Much of the catastrophe which he predicts is playing itself out around us now - he would only be surprised that it has taken this long (which is an interesting point to consider). Here he describes the meta-language (Platform for Change p.39):

Stereotypes are often useful concepts to deploy
Indeed it seems likely that we cannot manage without them.
What is wrong with the ones that we have
is that they relate to a vanished world.

In a changes world, the old stereotypes
give rise to undecidable propositions.

For example, if the means of livelihood
must be emoluments derived from work put in
to enterprises which themselves must make a profit
then

thinkers
teachers
artists
social workers
government

cannot be paid at all unless
everyone agrees to pretend
that the work they do is profitable
not in the general sense of the word
but as having a measurable monetary value.

But this remains a pretence.

The question : what are teachers, policemen, nurses worth?
is strictly undecidable within the language
of the stereotype.

That is why none of them is properly paid.

And as the world moves steadily away from the Homo Faber culture,
which it is doing because the production of goods
becomes more and more automated,
not only individual decisions but whole policies
become undecidable within the language.

The answer is not to invest bogus measures of benefit
but to devise a metalanguage
in which questions of value can be set and answered
in quantified but not monetary terms.

Note that this remark is meaningless
rather than merely puzzling
inside the existing language:
that language decreates the measures we need.

Any new argument must be competent to describe
new kinds of organization too.
The Argument for Change declared that such organisations
would look like 'anything organic'.
The passage as spoken entails the existence
of commonalities between all organic systems.

That there are such commonalities
is a fundamental tenet of cybernetic science
(as will be argued later).
The statement of commonality is necessarily metalinguistic.
Thus I see it as something more than a generalisation
of the machinery of particular systems -
for such a generalisation would be a perception
made by a neutral observer.

The metalanguage will be something more than this:
it will be positively generated at the focus of meaning
that lies beyond an entire range of undecidable propositions.

So the second comment is this.

We should try to envisage the developing general thgesis
in systemic terms
so as to show not only how things are connected
but also how the inherent undecidability
of the language used in each system
is expected to generate a metalanguage -
and of what form.

----------------------------------------------------

The idea that the metalanguage will be "positively generated at the focus of meaning that lies beyond an entire range of undecidable propositions" is intriguing. I want to say that the metalanguage will be positively produced by the constraints which limit decision-making in the old system. I want to say that those constraints autocatalyse new thinking. The problem is that they do this whilst at the same time destroying the conditions within which any new thinking can flourish.

@Telegram, Raspberry Pi, Block Chains and Educational Transactions

@Audreywatters is collecting some interesting stuff on block chain at the moment. I think one of the fundamental objections to Block Chain is that it is claimed that education is not transactional. Perhaps this assertion is made without thinking too much about what a 'transaction' is, or that it is as simple as 'buying and selling' (as in BitCoin) and that education is clearly more than that. But in economics, although buying and selling are the classic transactions, at a deeper level, they are quite mysterious intersubjective things. Von Mises and Hayek's focus on 'Catallactics' - the science of exchange - becomes very important here, I think. If Block Chains are about transactions, that means they are about catallaxy. Now what does that mean?? Perhaps we should distinguish between a kind of "potlatch catallaxy" and the "market catallaxy" that Hayek focused on. I think potlatch is at the heart of educational transactions: when education works, it is about giving.

The value of the block chain discussion is that we start to see transactions everywhere in education. I recently created a Raspberry-pi/Arduino-driven registration system for some medical students. They are employed by the health authority, who want twice-daily reports on their attendance. My hardware uses an NFC reader to scan their student cards, store it in a database (on the Pi) and provide the data on a simple web interface which the course administrator can query.

The student waving their card in front of the device is a kind of transaction: a statement "I'm here. I comply with the regulations." This transaction is of course the prelude to what the student ought to be doing next - listening attentively and learning how to be a medic (well, that's what we hope!). The students wanted some visual feedback from my registration scanner (at the moment it only beeps). I was thinking about interesting ways of doing this when I stumbled across the Telegram app (see https://telegram.org/).

Telegram is different. It's a communication app a bit like WhatsApp, but it makes itself highly configurable - both on its front-end (so you can create custom interfaces) and (most interestingly) at the back-end. Most powerfully, in the back-end you can program Bots. Bots can communicate with Telegram clients, and can be integrated with all kinds of devices. So, I thought, "let's get the register scanner to send a personalised message to Telegram when the students scan in". I'm working on this now... but having played with the Telegram API, wrestled with SSL certificates (it works through SSL), I'm now very excited about how the transaction of registering could be integrated with all kinds of other transactions - for example, the students' engagement with their e-portfolios, or real-time class tools like Socrative, or other physical devices. Or with each other. With Bot we can play games. It's possible to create group activities where individual students have to seek out other students to work on a problem together, or groups have to break apart, intermingle, go to different locations, collect data, interact with artefacts, and so on. Suddenly the combination of the ability to track transactions combined with the ability to capture different kinds of engagement in different kinds of environmental setting makes possible coordinations of learning across a range of environments which simply hasn't been possible until now. It's what the Alternative Reality Gaming people were aiming for - what things like Blast Theory's wonderful Ivy4ever (http://www.blasttheory.co.uk/projects/ivy4evr/) did using text messaging.

I think sophisticated transaction handling can create new forms of conviviality which do not depend on heavy-handed and unsophisticated management of the educational cohort. Given the mess that some managements are making of education at the moment, and the fact that the transactions they engage in have little to do with learning (what's this about, for example?  http://www.theboltonnews.co.uk/news/14255310.University_hits_back_at_claims_it_is_one_of__UK___s_least_transparent_/) I wonder at the possibility of replacing the whole of lot of them with computers. We certainly use computers in a very bad way in education: they should be used to automate management, not amplify existing practices.

And of course, there's really cool stuff we can do. My engagement with telegram came from this from Nick Johnstone (http://widdersh.in/controlling-sonic-pi-from-vim-or-anywhere-else/). This is why we should not give up on technology! The requirement is (as ever) for a combination of sensitivity, humility, creativity and reason - with some carefully considered political steering.

Educational Transactions, Artistic Transactions and Improvisation

One of the things that fascinates me about Live Coding (see http://dailyimprovisation.blogspot.co.uk/2015/06/improvisation-and-performing-coding.html) is that every improvisational act is made explicit - there as a record of a particular aesthetic decision taken at a particular point in time, in a particular context. As I hack around with Sonic Pi and the equally wonderful and more sophisticated Overtone this transactional aspect of improvisatory experience in Live Coding is becoming more important to me.

This is at a time when I am exploring (in a potential bid) the possibility that education itself might become more transactional in the same kind of way. I've had many arguments with friends about whether education is or is not transactional. I say that as far as any kind of final assessment is concerned, of course there is a transaction there: "I've done this work; you give me a mark". The problem is that this transaction is large-scale, processed in batch and only done once at the end of a learning process which has involved many much smaller interactions which are also transactional and which are never captured, nor can ever be inspected.

Education is organised in the way it is because it cannot break its transactions down. It has enlisted technologies to reinforce its traditional batch-mode of operation, and reinforce the status of institutions which are seen as the guardians of 'quality' - a substitute for the inability of the system to expose its decision-making processes. So the inability to handle transactions has implications not just in the management of learners on an everyday basis, but the global organisation of education, access to opportunity, social status, elitism and so on.

Coming back to music, and seeing improvisatory music as a set of artistic transactions, in a group there are processes of evaluation and ranking of individual decisions. One of the things which I am about to explore is group improvisation using Live Coding. Here, individuals not only listen to what they do, but to what others do. They copy. In doing so, a natural structure of ideas and the creative agencies which produce those ideas emerges.

Teaching and learning is like this too. Lecturing, for example, is usually quite an improvisatory performance: good preparation for a lecture means leaving enough space to react to the situation among the learners. Teachers and learners make declarations about concepts, or rather reinforce other peoples' declarations. Teachers make declarations about the declarations made by learners (right or wrong...) What if this was captured in the way that Live Coding captures a musical performance? I'm not suggesting that teachers and learners actually engage in writing code as part of their interaction, or even that a big brother system captures every utterance. But we lose consciousness of these transactions because we've got so used to the assessment (the big transaction) being the only thing that matters, we've forgotten to listen to the small transactions. At least acknowledging the transactional nature of what we do would be a good start.

First of all, it would discourage teachers talking too much. It would encourage more "What do you think?" moments. And it emphasises our inability to capture the What do you think moments and stand back from them and reassess our approach. Maybe some real-time polling systems can be used for this (e.g. Socrative) - personally, I'm much more interested in telegram.

There's something else about recording transactions which I think is important. Education operates in batch mode partly because of the conviviality of the classroom. The class we-relation is important - even if the lecture is terrible. What is that we-relation? Alfred Schutz, who invented the term, thought that it had something to do with the shared passing of time. Music creates a we-relation even when the performers are no co-present with the listener, or the composer is dead. Might a transaction record similarly have this temporal quality? There's a need to experiment...

Sociology and Lakatos

There was an interesting twitter exchange today between @profstevefuller and @mark_carrigan about sociology: https://twitter.com/ProfSteveFuller/status/707860253682110464. Fuller stated that

"So-called 'public sociology' often looks like a kind of ventriloquism in which the sociologist is the dummy."

This questions not only the scientific content of sociology, but also its mission. I found myself , by coincidence, reading Lakatos's "The methodology of scientific research programmes" today, and found him make this statement (I leave out his discussion on Newton and Halley for brevity - but it is worth reading)

"all the research programmes I admire have one characteristic in common. They all predict novel facts, facts which had been either undreamt of, or have indeed been contradicted by previous or rival programmes. [...] Einstein's [research] programme [...] made the stunning  prediction that if one measures the distance between two stars in the night and if one measures the distance between them during the day (when they are visible during an eclipse of the sun), the two measurements will be different. Nobody had thought to make such an observation before Einstein's programme. Thus in a progressive research programme, theory leads to the discovery of hitherto unknown novel facts. In degenerating programmes, however, theories are fabricated only in order to accommodate known facts. Has, for instance, Marxism ever predicted a stunning novel fact successfully? Never! It has some famous unsuccessful predictions[...]"

I think the attack on Marxism is a broader assault on sociology doing what Fuller accuses it of. The ventriloquism is the process of making the theory (the sociologist's utterances) fit the facts (the agent's utterances).

The point about the prediction of novel facts is not, I think, about any kind of superiority of physics over sociology. It is about the demarcated roles of theory vs. experiment. From a cybernetic perspective it was quite simply expressed by Ashby: theories generate logical possibilities; experiments tell us how those logical possibilities are constrained in nature. The tension for any scientist is between explaining things, and identifying constraints or errors that are revealed through exploring the logical possibilities. It is, however, from the identification of error between theory and nature that science progresses. But errors exist at multiple and entangled levels: matter, bodies, discourse, ethics and so on.

The discovery of constraints ought to lead to theoretical refinement, which in turn ought to lead to new generative possibilities. It is this last bit which Marxism hasn't done properly: it refines its theories to fit the facts, but then doesn't re-run the implications of the adjusted theory. It simply tries to patch itself up. Lakatos puts it:

"Marxists explained all their failures: they explained the rising living standards of the working class by devising a theory of imperialism; they even explained why the first socialist revolution occurred in industrially backward Russia. They 'explained' Berlin 1953, Budapest, 1956, Prague 1968. They 'explained' the Russian-Chinese conflict. But their auxiliary hypotheses were all cooked up after the event to protect Marxian theory from the facts." 

What happens with this patched-up social theory is that it becomes so open to interpretation, the advantage of having a codified theory as an object to coordinate scientific engagement is completely lost. There is no agreement about theoretical categories, and no agreement about empirical procedures. Particularly without the latter, the discourse floats away - it has nothing to tether itself to: there is no shared object in the lifeworld around which scientists can have a constructive discussion.

The importance of any theory is its codified logical consistency which is the source of possibilities which may or may not actually exist. These are the source of new distinctions to be made about measurement (like Einstein's distance between stars), or the constraints that are discovered by a social theory not fitting lead to refinement which generates new possibilities amongst which might be some new insight into a "novel fact".

My favourite piece of predictive social science would probably be sniffed at by many sociologists: Winograd and Flores's "Understanding Computers and Cognition" of 1986. Not only is there a refutation of the prevailing view of technology at the time, but a clear and accurate prediction of where things would be 6 years later (and 6 years is a long time in technology!)

Musical Ecologies

I've got my head in some pretty technical stuff at the moment which I find exciting and energising, but which, as with all technology, is likely also to be ultimately disappointing (although the energy from it will move me on). I'm used to this pattern: I go through phases of functionalism where I am fascinated and motivated by what might be changed through technology. Whilst I'm in it, though, I also feel the need to temper my enthusiasm in some way.

Music is my corrective. I've been thinking back to this: http://dailyimprovisation.blogspot.co.uk/2015/09/entropy-and-aesthetics-some-musical.html. It's about music and ecology.

When we design technical systems - even technical systems which are meant to be emancipatory for people - really we are designing something with ourselves as designer in the driving seat. The status function for the technology is made by the technical designer/hacker. I've been making a lot of status function declarations about blockchain recently. I think it's important because it's certainly a new kind of thing, and its implications need to be investigated. But in making those status function declarations, I put myself in a privileged position. Do I wish for some social acclaim which says "Ah yes! Johnson was right all along!"? Wanting to be right is a real problem (I think I am right about that)

A piece of music is a kind of declaration. But it's a declaration that comes from the heart as well as from the pen. The heart-bit is important. Feelings are communicated in music because, I think, we share a sense of time with the performers and other listeners. This despite the fact that many composers are long dead. It still communicates in what Alfred Schutz would call a pure 'we-relation': i.e. a form of communication which is usually only available in intimate face-to-face exchange. This is unlike most declarations, including declarations about technology. They are made in the 'world of contemporaries' - a world of remote relationships with people who live at the same time as us, but who we do not share a 'vivid simultaneity'. Indeed, we have very little understanding of what our technological propositions might do to them.

When we try and pursuade someone face-to-face about the value of a technology, we find a way of communicating face-to-face which bypasses the deeper emotional aspects of the pure we-relation. Technologies turn our face-to-face engagements into remote engagements which occur face-to-face. This is the great trick of capitalism: the bypassing of time and intimacy, and its replacement with terminology codified in the world of contemporaries. Too much academic engagement in Universities has become like this too. Psychotherpeutic, curiosity-driven learning rarely happens: instead we drill learning outcomes.

Coming back to music, we can codify its parts in an abstract way (again, the world of contemporaries). But in shared experience in time, we can see that none of these parts are separable, any more than the organs of the body are separable. Moreover, there is no executive function among these parts: each part relates to the others in a fundamental, intertwined way. Just as the brain does not 'control' the body, but ensures that effective relationships are managed between all its components, including itself. 'Management' is inherent in the organisation of music.

That is what I have to ensure in my thinking about technology. The management must be inherent in the organisation of the technological components. In educational technology this is a real challenge. Only if this is achieved will a technical system start to feel like music rather than an authoritarian state.

Measuring the effectiveness of the organisation of music, or technology, may be the key thing. These, by necessity, are measures of ecological viability.

Searle's Status Functions and the Educational Block Chain

@dkernohan made an interesting blog post about blockchain yesterday (http://followersoftheapocalyp.se/the-chain/) which also contains some great links to current work and thinking. When new technologies come along, particularly in a time of great confusion in the world - not just about education, but about everything - it's exciting for some. I'm one of the people who gets excited about technology. This is partly because technical change is so often the herald of social change - and the prospect of change is a sign of hope that the world won't continue to be as terrible as it currently is. However, technology frequently disappoints. People either don't use it (which may in fact be the best outcome!), or if they do, managers will exploit the fact that the tools work to exert greater control on everyone else. In recent years, technologies which promised emancipation, disruption, subversion and autonomy appear to have accelerated and exacerbated inequality, surveillance, control and oppression. Yet I still find the prospect of technological transformation exciting.

I think this is partly due to the nature of the functionalism which surrounds any technological development. Functionalism is inherently positive - a can-do, "we can fix it" state of mind. Part of the problem with functionalism is that it loses sight of how power operates, and is often blind to the power inherent in its own propositions.

Recently, I've begun to see power through the perspective of John Searle's idea of 'status function'. A powerful person or organisation is something capable of making a assertion about the existence about other social entities, and to have that assertion supported by the general 'collective intentionality' (Searle's term for the common thoughts of a community) to support it. They have what Searle calls 'deontic power' to make a 'status function declaration' - in other words the assertion that a banknote is a banknote, a VLE is a VLE, a degree is a degree and so on. Technologists make status function declarations about technology. If nobody believes them then the technology doesn't take-off. If powerful institutions believe them, and support the status declaration about a technology, then the technology is more likely to be successful.

Entities in society which have the greatest deontic power tend to be institutions like governments, banks, universities, monarchs, priests and the media. It is interesting to reflect that in recent years, each of these institutions (apart from the universities - it's only a matter of time) has had its deontic power questioned by the exposure of scandals.

Interesting things happen when status function declarations get mixed up in unusual ways. That's basically what's happened with BitCoin. BitCoin plays with "collective intentionality" in a new kind of way. Technologists make a status function declaration about the trustworthiness of the "distributed ledger" which underpins BitCoin. There is a vested-interest among the investors in BitCoin to uphold this status function declaration. Not just because they like the idea of a currency without a central bank, but because in supporting the status function declaration about the ledger, they support the value of the BitCoin currency which they possess. The dynamic of trust (the collective intentionality), and the dynamic of reward is finely balanced - and the reward for the individual reinforces their desire to avoid the centralisation of institutions.

This point is very important when we consider the potential of BitCoin in areas like education.

It's not enough simply to declare the existence of what amounts to a kind of distributed database for educational transactions. Such a declaration alone will have no power, because no institution will support it, and there is no loop of self-interest by its users which can subvert institutional authority. It will be too difficult to get people to submit their data to it: what do they get out of it?

Somehow the loop has to be closed. What matters in education is the declaration of status of learners: the gradual acquisition of their own 'deontic power' (I am qualified to say such-and-such because I have a degree from...) So collecting transactional data is not enough. There needs also to be a mechanism for validating, comparing and ranking between individuals.

So what if we had a process whereby on the one hand, educational transactions are recorded, but alongside it, the emergent transaction ledgers are validated and ranked? And what if the process of validating and ranking educational transaction ledgers reinforced ones own educational transaction ledger (so a judgement about someone else is a transaction). Would that close the loop?

My incentive for recording transactions would be to enable me to make judgements about the transactions of others, which would enhance my own ledger of transactions. Is this a closed, self-organising non-institutional status-enhancing machine? Possibly. But one thing seems clear to me - the chain needs to be a loop.