ARPA, ARIA, defence, and innovation: when research goes to war

Innovation has both a history and a mythology.

Or, more accurately, it has several. Our choice of narrative tells us a lot about what it is we want to support. Dominic Cummings, for instance, cleaves strongly to the US (Defense) Advanced Research Projects Agency ((D)ARPA) model – a story about freedom, nurture, and happenstance. And, as it happens, a myth rather than a history.

The use of innovation, science, and technology as ideas to make the development of larger and more complex military projects palatable to the public continues, of course. As of today it is UK government policy to “incorporate science and technology as an integral element of our national security and international policy”. The government is developing its own science and technology horizon scanning and assessment capacity, and £6.6bn of research funding will flow through the Ministry of Defence. Universities don’t get a mention.

Research and development success is, once again, a matter of national security.

ARPA in context

The argument, as codified into the Advanced Research and Invention Bill, is for long-term support for wide ranging science and technology research and the near abolition of bureaucracy. This almost entirely fails to describe the historical (D)ARPA – it was actually established as a bureaucratic mechanism to avoid military bickering about the home of the nascent space programme, holding this role for around one year before the establishment of NASA. The sole organisational innovation was the ability to award contracts quickly using military processes more normally used in wartime – a practice that led to Operation Argus: the atmospheric explosion of a 1.7kt nuclear warhead over the South Atlantic in an attempt to generate a “forcefield” of charged particles to protect against nuclear weapons (It didn’t work).

It’s impossible to think about the early years of the space programme without that jolt of idealism and clear-eyed optimism that the modern breed of innovation often lacks, but in many ways the space programme was cover for a huge investment in the cold war arms race – ARPA’s role was really in the development of novel weapons and approaches to war. When space research and development moved to NASA, ARPA pivoted to the development of counter-insurgency techniques that arguably caused – or at least intensified significantly – the Vietnam War (you can read more of this in Sharon Weinberger’s superb The Imagineers of War).

ARPA did make some notable discoveries that had lasting impact in pursuit of their military aims – the approach that underpins GPS saw development via a means to track Russian satellites, the legendary ARPANET (which pioneered the packet-switching approach to networking that underpins the modern internet) was derived from a means to route firing messages when parts of the armed services had been destroyed. Serious investment in the science of seismography stemmed from a need to monitor nuclear testing. But there was no model of innovation that drove their development, and other vast swathes of funding (on counter-insurgency, and on mind control) produced less than nothing.

I hope that the intellectual problem is coming across alongside my long-held pacifism here – randomly funding a bunch of projects based on little more than whims is not the best way to drive innovation. Universities and academics did contribute a great deal – the smell of money can tempt any university – but very much as resources rather than as leading curiosity driven research. If we can only fund network theory or geophysics by politicising and militarising it, we have a more fundamental problem with our theory of change.

The imitation game

A peculiarity of ARIA is the decision to promote a US model of state support for innovation during a time where British exceptionalism is in the ascendency. Doesn’t the UK have a foundation myth of innovation?

It does, and like all of our foundation myths seem to, it comes from the second world war. Bletchley Park was the birthplace of Tommy Flowers’ legendary “Colossus” – the world’s first programmable digital computers, and was the scene of important wider work on radar, signals intelligence, and computing. Why, we may fairly ask, do fans of organisational innovation never cite Bletchley Park as an example?

The answer is as simple as it is startling – as an organisation Bletchley Park was shambolic. Christopher Grey’s seminal Decoding Organization: Bletchley Park, Codebreaking, and Organization Studies tells the story of a number of overlapping organisational structures in an environment riven with what I can only describe as office politics. Following the 1941 Van Cutsem report (very much a wartime Ofsted inspection that put the place into special measures”), major reorganisations in 1942 and 1944 attempted to solve some of the many conflicts between services and civilians, code-breakers and translators, and a number of distinct and non-complementary management hierarchies.

These reorganisations did not facilitate or grow innovative practices at Bletchley – the Lorenz cipher was broken in 1941, the refinements to the original Polish military Enigma break made in 1939-40 – though they clearly made the facility an easier place to work and work with. But world-changing innovation flourished during a period that could best be described as chaotic, unstructured, and under-resourced.

Benign neglect and active promotion

ARPA’s pivot to counter-insurgency and computer networking could be seen as the actions of an agency that had a budget but no direct political interest. Bletchley had a political goal but very little day-to-day input, and was in competition with other military interests for a scarce budget. Neither, in other words, was designed or encouraged in developing innovation – both did.

Both stories have been used to promote commercial and industrial innovation as a government priority. In the UK, straight after the war, the Barlow and Percy reports attempted to design and implement systems that would support scientific research and scientists. CP Snow’s now unfashionable 1959 Rede lecture on The Two Cultures set in motion a national conversation about the place and value of science that led indirectly to Harold Wilson’s famous “white heat of technology”, which described both an economic and societal benefit from direct government involvement in innovation.

In industry, the early work of Schumpter on invention, innovation, and infusion became a common frame of reference as research and development departments were systemised and the question of the appropriate level investment in such processes became a national and global one. The literature on innovation took a turn into market-focused thinking in the latter part of the twentieth century – ideas of the niche and design dominance placed competition at the forefront of a race for innovation. In the last two decades models of iterative customer-driven innovation (exemplified by the group of approaches that form a part of the “agile” ecosystem) and Christensen’s now largely-discredited concept of “disruption” situate innovation far closer to product development than basic science.

In this historical conceptualisation (which I admit is crudely drawn) you could see the swing of the pendulum back to the “moonshot” models that delight both Cummings and Mariana Mazzucato as a necessary corrective to a market that has not given us the competitive innovation we hoped for but has given us innumerable iOS apps.

Big innovation in government

Moonshots do bring ancillary benefits, but to work we need to be very clear that we are indeed aiming specifically for the moon. And researchers and scientists do not get to make that choice of destination, or even (as the story of ARPA tells us) decide the other things we see along the way. Only the president got to point at the moon – only the defence establishment got to choose a plausible cover for missile development.

We should probably be more startled than we are that a conservative government is making the case for “picking winners” in this way. ARIA’s founding bill leaves the field open for government appointees rather than academic peers in selecting and funding projects and in abruptly terminating unpromising work. Unlike ARPA, ARIA will not have a defence remit, unlike Bletchley there will be clear chains of command up to (Prime?) Ministerial level.

It will bring more control and less accountability to government interventions into what is drawn as in intervention in the private sphere of commercial innovation – effectively situating the wide end of the innovation funnel as a state good but without any of the benefits (an innovation commons, diversity of thinking, a global rather than UK perspective) that academic institutions bring.

Let’s talk about universities

You could see use of universities as research funding organisations as kind of a mini-ARPA model: funding is made available for specific projects, for wide-ranging basic research, and for commercialising research based on previous performance – but funders never know quite what will come out of the other end. Funding bodies can allocate money to research performing organisations (or parts of them) very quickly using standard agreements. Funding can be project based, post based, or (in the case of QR) open-ended.

Politely, the organisational model of the university is closer to Bletchley than ARPA. This does not stop innovation and development (most recently, the development of vaccines for Covid-19) – neither, to speak frankly, does it make universities a particularly nice place to work.

Culturally, universities are in a bad space – despite being able to carry out all this amazing research and development while upskilling the wider population and being one of the largest export industries in the country. Disagreements about positioning in a “war on woke” make another research support method politically necessary and ideologically expedient.

We cannot celebrate the recent growth in research funding without thinking about the cause of it. Whatever you feel about the military use of research, it is once again becoming a priority. Considering the models and processes that do and will underpin research shows how essential accountability and transparency really is.