This post writes up a short thread on Twitter on memetics and economics. One that tells the story of a ‘paper’ I wrote perhaps 15 years ago or more, now, with the same title as this post, and sent to a now defunct journal ‘The Journal of Memetics’. I don’t know why the journal folded. Discussion of memetics is curiously absent on my Twitter feed. Perhaps the idea proved a dead end, or just died for contingent reasons. The rise and fall of memetics seems beautifully recursive as a phenomenon, given what memetics is.
You might well conclude from reading this that I entered a period of temporary insanity; or that I have more heterodox tastes in economics and social science than you’d up to this point guessed from my blog and feed.
It’s also possible that the connection explored here was made before I first thought about this in the early 2000s. Or since. I don’t read much heterodox economic thinking, and have not tried to trace where memetics went. In these interstices of intellectual life there may reside accounts of the thoughts that are set out below. I should note that Leigh Caldwell responded to my thread of tweets saying that he had been writing about taking the perspective that beliefs themselves, and not the individuals that hold them, should be the unit of analysis. So credit – or ridicule! – perhaps should go to him too. And, as you will see, the ideas here are really intrinsic to the original accounts of memetics, so if there is any contribution here, it is very small.
What is memetics / what are memes?
Memetics was the term used by Richard Dawkins and, later, Susan Blackmore and others to describe the topic of analysing ‘memes’. ‘Meme’ was the word used by Dawkins in The Selfish Gene to describe an idea, or concept. The word ‘meme’ itself, ironically, mutated to be used more familiarly as ‘internet joke’. Here, I am using the original meaning. Internet jokes are a subset of all ideas. [At least, it may yet come to pass, but so far there are ideas that are not internet jokes]. So the mutation involved new memes referring to a subset of old memes.
Dawkins’ original insight or proposition was that one could view ideas and concepts as we view viruses or parasites in bodies. Viruses and parasites are to be taken as pursuing their own interests, maximizing the chance of perpetuation, and they do this by invading hosts’ bodies, and manipulating those bodies in pursuit of their own objectives. If, like me, you have enjoyed readng popular biology books, you will have encountered ghoulish accounts of parasites that hijack insect bodies to put them in the way of their prey, the objective being that the parasite entires the insect predator. But more prosaic examples are of course the common cold sneeze. Which, on the face of it, looks like a mechanism to get the host to project the virus into the air and towards the next host.
Suppose we hear someone humming a tune. You might analyze this in terms of an individual taking pleasure from re-creating the music for themselves, either because they liked the music, or because they like experiencing their facility in humming, or are practicing for some future singing exploit that they find pleasurable. But you could also take a memetic perspective. This involves taking what [I think] Daniel Dennett called ‘the intentional stance’ towards the tune itself. Taking that stance involves conjecturing that the tune can itself be taken to have intentions – in this case maximizing the chance that it propagates around the population of human minds capable of recycling it and perpetuating it. This is of course not to say that the tune has concscious or even unconscious ‘intents’ any more than a virus does. But that the process of natural selection will inexorably have shaped tunes, just like viruses, to be such as to be describable as if they did have these intents.
The notion that memes can be thought of as like genes requires that there is natural selection operating on memes. Natural selection in genes requires: 1) copying, 2) errors in copying and 3) pressure in favour of ‘good’ copying errors [‘good’ here referring to the chance of making it to the next act of copying]. Is there natural selection in memes? There is copying: person A invents a tune, plays it, person B hears it, has the compulsion to repeat it. There is fallible copying: persons B to F might remember the tune perfectly, but person G makes a slight error. And there is selection: we can’t remember all tunes. Only the most memorable or catchy, or compulsive or pleasurable get retained. The rest, like ‘i before e’ on Yazoo’s albumn ‘Upstairs at Erics’ get forgotten by almost everyone.
What has memetics got to do with economics?
What has this got to do with economics? Most you sufficiently interested to read on to this point, and those who have come across memetics/memes and economics before, are likely to have jumped ahead to the conclusions already. But since economics is a subset of the study of human behaviour, and if memetics potentially explains all human behaviour, then memetics potentially explains all economic behaviour. The workhorse model of economics often carries implications for policy design; and moralistic reasoning often flows in the minds of some from conventional analyses of economic behaviour. Ditching this workhorse model therefore necessarily has implications for policy design and moral thinking about behaviour.
Economics and rationality and memes
The working-horse model for economics, often referred to disparagingly as ‘homo economicus’ conceives a human as a single coherent mental entity maximizing utility; and firms as to be thought of as coherent single entities maximizing profits [a corollarly of them acting on behalf of shareholder-consumers maximizing utility]. There are exceptions and complications to this, some of which I will come to. But for now ignore them, so we can make the point clear.
Memetics implies the proposition that, to some degree, either partial or total, these assumptions might be false. Behaviours – according to this theory – are not manifestations of utility maximization of the brains that instigate them any more than the sneeze helps me when I am infected with the common cold virus. These behaviours are maximizing an objective peculiar to themselves. Which, to re-emphasize, is simply their propagation into the next round of behaviour.
A conventional model of the difference in savings between one country and another, or between one individual and another might posit a difference in the tilt of the expected income trajectories of the two agents. A young person expecting to earn more tomorrow than today would borrow, bringing forward consumption from when it would otherwise be abundant [and increase pleasure less at the margin] to when it is scarce today [and therefore increases pleasure more at the margin]. A memetic explanation of differential savings behaviour would say simply that in one country the meme for saving a lot had spread, and in the other, it hadn’t.
Similarly, rationality is invoked to explain: behaviour – including portfolio selection and asset pricing, labour market choices – in response to risk; labour suppy and unemployment in the presence of taxes and benefits; behaviour in response to safety or public health regulations, like seatbelt wearing or sugar consumption. In all these cases, we can at least conjecture that memetics, not conventional rationality, is behind what people are doing. Risk lovers just got the risk loving meme. Workaholics are not responding to taxes, but they got the workaholic meme. Sugar eaters got the don’t care less about government health warnings meme. One could think of hundreds of other examples, many more convincing than these.
Policymaking in the face of rational behaviour often invokes the logic of incentives. For example: Keep benefits low to stop people idling in unemployment. Fine people who don’t wear seatbelts. But if incentives are not at work, or not wholly at work, and memes are behind behaviour instead, or exerting an influence at least, then policies might not work as anticipated.
Moralistic reasoning about behaviour also often follows from thinking about behaviour in terms of rationality [not just rationality of course – more later]. For example, a common view on the right is that people on benefits are lazy. Translated into the language of homo economicus, they place a high value on leisure, and would need benefits to be very low, or the real wage to be very high, to induce them to work. This kind of moralistic reasoning falls away with memetics. Although the version suggested here does not seem appealling either. That would be ‘people on benefits have been hijacked by the meme for not working’. There might be others, of course: they could have been hijacked by the meme that there is no work to be found when this is not the case. And we could think of more possibilities [it is going to be a problem for this theoretical view of things that there are lots of possibilities].
Of course moralistic reasoning can be demolished through other routes besides memetics. People don’t choose their endowed abilities, upbringing, or other draws of chance in the sequential lottery of life.
Economics and models of non-rationality
Economics is not just about analyzing behaviour through the lens of perfect optimization, of course, although the meme that it is precisely just that has spread far and wide, and is often used as a joke about our profession. Famously, Nobels were awarded to Kahneman and Tversky for their research into what was dubbed ‘behavioural economics’. The many ways in which our behaviour seems not to match up to rationality. But be comprised of a hotch potch of short-cuts, biases, heuristics and misapprehensions. Connections are drawn to evolutionary psychology, which again, contrary to the meme [see what I did there?] that this is about thinking of every feature of the brain as perfectly adapted to its purpose, is full of work examining the engineering and perception constraints of cognition; or the gap between behaviour or behavioural compromises that suited the environment which shaped our evolution, pre civilization, and the one we operate in now.
Memetics and behavioural econ
Memetics has just as destructive implications for this way of thinking about human behaviour as the more old fashioned rational homo economicus. Behavioural econ still posits the individual as a single coherent entity, albeit one operating subject to biases, cognitive and engineering compromises, and the like. Memetics, in the extreme, [repeat: you don’t have to swallow that memetics explains everything, all the time, everywhere, to be a valid contribution to explaning what we see], sees all behaviours as manifestations of memes for those behaviours, which are suiting themselves. Evolved cognitive biases may be overcome, but will never go away. Memes can be forgotten, or replaced by others.
Rather obviously, just as the appropriateness of policymaking via incentives in the presence of rational economic humans is questionable if instead memetics is at work, so too policymaking that follows from deducing behavioural biases are at work may no longer be appropriate. For example, ‘Nudging’ might never work. Or if it does, not always be necessary, or enough.
You don’t have to think hard about this to realize that if what I have set out thus far follows, then memetics has implications for what one reads into conventional econometric analysis.
Studies of the labour supply responses to tax changes that attempt to draw inferences about preferences regarding work and leisure are searching for something that doesn’t exist at all [in the extreme version of the memetic account of behaviour] or are in a tussle, potentially time-varying, with memetic forces [or ‘memes’].
Estimates of the RBC or NK model from aggregate time series that attempt to make inferences about parameters in structural relations [like the IS curve, or the structural Philips Curve[s]] that have their origins in assumptions of consumer and producer optimization are analogously going after something that doesn’t exist, or is not time-invariant. If this theory has anything to say.
The natural question arises, then: how would we know if the theory has anything to say?
I mentioned earlier that a problem for memetics is that there are lots of possibilities. In principle, everything you see anyone and everyone do can be posited as the manifestation of the temporary infection of the brain of the person doing it by a meme. If two behaviours differ, that is to do with differential meme infection. If behaviour changes over time [alarm bell – the equivalent of ‘preference shocks’ in RBC/NK analysis!] this is to do with fluctuating influences of a set of memes, or changes in the memes that win out.
There is an inherent unfalsifiability about memetics generally – and therefore in the proposition that economic behaviour is memetic. How you respond to that is the subject of an old debate in the philosophy of science. One response is to decide that theories like this are useless and move on. My own feeling is that unfalisfiability may just be tough luck. You can move on to some other mode of guiding policy if you like, but if memetics happens to be true, you had better beware that that other mode may not help.
It’s also worth noting that optimization based theories of behaviour also suffer from an inherent unfalsifiability. Evolutionary psychology might start from a conjecture that some behaviour / brain features is the optimal response to a problem. Rational economic human models likewise can [often do] simply proceed by assuming what we see is the output of optimization, and confines itself to finding out what constraints are operating that the agents are optimizing in the presence of.
Just as with conventional modelling based on rationality – not only in economics, but in evolutionary biology – it is possible to make progress. If you can collect data on the technology and constraints that face an agent, you can then test assumptions about rationality. Perhaps likewise, if one could say more about and measure the factors that allow or prevent memes to spread, one could estimate the importance of memetic influences in behaviour.
One final joke
I wrote this paper up and sent it to the Journal of Memetics. At least I think that was its name. My memory might be playing tricks. And it seems inappropriate, given the subject matter, to google and check.
I got a friendly revise and resubmit from the editor, called Paul someone. I forget his surname [insert repeat of above joke].
Unfortunately, either because I thought there was not much incentive for someone who wanted to signal mainstream credentials, or infected by the mainstream meme, I took a long time getting round to it. After that, the laptop it sat on was stolen, and then the journal folded. On such slender chances did the fate of the meme that economics may be a subset of memetics hang. Of course, if anyone out there edits a journal they think might want a brief account of these thoughts, I’d be happy to submit again.
Ironically, the only economist I ever discussed this with before was Mark Dean, a former colleague at the Bank of England who went on to do a PhD at NYY and become a celebrated pioneer of neuroeconomics. Neuroeconomics exploits advances in brain science and technologies of analysing brains to look for the biological wirings that correspond to or refute the axioms of modern economics. Memetics, potentially, affects neuroeconomic research methods and conclusions. It is no use looking for the hard wiring of risk or leisure preferences, if brains are just the vassals for a fluctuating infection by memes. Inferences about apparent fluctuations in preferences [there are experiments deducing just this] under a memetic interpretation would be evidence of those varying infections.