Friday, January 14, 2022

 537 Certainty and uncertainty

 Certainty is scarce, though it does exist. There can be certainty of something when it logically follows from the axioms of an underlying mathematical system that you take for granted, but it is conditional on that. That is called ‘analytic’truth, by deduction, in contrast with ‘synthetic’, empirical truth.

Uncertainty can be denied, by religion, with a benevolent God, belief in an afterlife, and may be hidden in ritual. Natural scientists used to search for the certainty of laws of nature. David Hume warned that no matter how long one has observed a regularity, this does not logically imply that it will continue in the future. The regularity of nature is an assumption. One needs a causal account to justify the regularity of any phenomenon. This inspired Kant to the idea that we do not know reality as it is ‘in itself’, but only by the working of mental frames of interpretation. I would give that a little twist in adding that one does not know to what extent ideas form and distort reality. If we do not know reality as it is in itself, how do we know that and in how far ideas misrepresent it? We can only have confidence in ideas to the extent that they satisfy logic, cohere with other ideas that have withstood the test of time and have survived a variety of criticisms.

Thoughts are formed in ‘mind frames’ that are ‘triggered’ by circumstance. An example is the following. Show someone a white napkin, a white shirt, a white sheet of paper and ask what cows drink. Often, triggered by the whiteness of things, people say ‘milk’, but the right answer is ‘water’.

On the other side of the dilemma, the view is that uncertainty is pervasive and inevitable. Physicists have become accustomed to the Heisenberg uncertainty principle, that one cannot with certainty measure both the position and the momentum of an elementary particle such as an electron. An electron circling the nucleus of an atom has a ‘cloud’ of probabilities of location. When colliding with something else, the cloud collapses to a determinate location. Nature was before seen to consist of things existing independently but is now seen as inherently conditional upon interaction with other particles. That yields a process view of things (Nooteboom 2021).

According to Karl Popper, theories can be falsified but not proven, and logically this is an incontrovertible claim Theory seeks laws that claim universal validity: given conditions A, B. always occurs. Popper proposed that the ethic of science is to search for falsifications. Others objected that it is not rational for scientists to do this. They have an interest in gaining recognition by theories that are confirmed, giving a rational incentive not to search for falsificationn, or even hide it when inadvertently found. In a famous debate (Lakatos and Musgrave, 1970), Popper admitted this and even pleaded for a certain conservatism of neglecting falsification, not only because such conduct is a fact of life, but also because it is rational in yielding information on the limits, the ‘true strength’ of a theory, and yields hints in what direction to seek renewal, in tune with Hegel’s principle of learning by failure.

Economists distinguish between risk and uncertainty. In risk, one does not know what will happen, but one knows what can happen. You can see this as a set of all possible occurrences or outcomes. One can assign probabilities to them and calculate ‘expected value’. If one does not know objective probabilities, the procedure is to assign equal probability to all possibilities, and adapt them as experience accrues concerning their frequency of occurrence. Under uncertainty, by contrast, one does not know all that can happen, but one can imagine some possible futures, in narratives about it, called scenarios. One cannot optimise a policy across all possible but unknown futures, but one can consider the robustness of a phenomenon or policy across a few such possible futures, choosing a policy that is not necessarily optimal in any of those futures but yields a satisfactory outcome in all of them.

Uncertainty is unsettling to many people, who are then tempted to deny the uncertainty of a phenomenon, or even the very existence of uncertainty, and to construe myths that do yield an imagined certainty, in conspiracy theories that yield apparent certainty, often encompassing a multitude of phenomena, for the sake of simplicity, while rejecting criticism, theoretical or empirical, as exemplifying the bias and interest of the conspirators, and thereby confirming rather than falsifying the conspiracy theory.

Economists seldom take uncertainty seriously, and mostly opt for treating it as risk, because under uncertainty they cannot engage in the calculation of an optimum, which belongs to the core of economics (Hodgson 2019). As a result, they cannot offer an adequate account of innovation or learning, where uncertainty is of the essence. With a treatment in terms of risk, they can manage to reconstruct innovation that is incremental, staying within the basic principles of established practice. In government innovation policy this is favoured, because one likes to plan innovation in concert with incumbent business leaders, politically necessary for acceptance, while still waving the fashionable flag of innovation. For the leaders of large business, with a stake in established practice, incremental innovation is preferable because it does not yield the creative destruction of more radical innovation crossing the limits of established practice and the investments sunk in it. Radical innovation is erroneously expected to follow from specifying intended outcomes and timetables of achieving them, which is a contradiction in terms. Radical innovation yields unexpected outcomes that are not yet useful now, which are then labelled as ‘failures’, while even in their failure they are useful in showing which avenues do not yield results and indicate new avenues to try. Furthermore, radical innovations that do yield a useful outcome have a large and wide impact that compensates for  the host of efforts that did not have impact. This yields a criticism by many researchers of policy where fundamental research is challenged to be useful in the short term, while its often large impact emerges in the longer term. This criticism was included in a report to the Dutch government that was developed under my supervision. It was not appreciated, for threatening to ruin the game of calling incremental improvement innovation.

An alternative perspective, accepting uncertainty, is that of evolution, which does not presume a rational ‘intelligent design’, but explains development of novelty as the outcome of a process of selection by the environment, in combination with the generation of variety submitted to selection, in nature by chromosome cross-over in sexual reproduction, mutation of genes and variations in the interactions of genes and their ‘expression’ under influence of their environment. In heterodox, evolutionary economics, variety is created by invention and innovation, by entrepreneurs, and selection is done by the market and institutions.

In science, research proposals and papers are regularly rejected by scientific journals because they do not fit in the core of the research programme adhered to by the editors and reviewers of the journal. This is an example of selection, in an evolutionary theory of science. However, scientists held back by this selection frequently create their own proprietary journal, thus creating their own selection environment. This also happens to some extent in nature, when organisms defend themselves against selection, or create a benevolent selection environment, with beavers building dams, rabbits digging warrens, or species engaging in symbiosis with others or being parasites. This matching of selection environments with life forms to be selected, is called ‘co-evolution’ When this prevents a stable selection environment, evolution breaks down. For libertarians who disregard market imperfections, this yields an argument not to interfere with the selection mechanism of the market.

An example is the present breakdown of selection of opinions by arguments of truth, logic and facts, since truth has been made irrelevant by demagogues and some governments, in the dissemination of ‘fake news’. Opinions are now presented as certain, incontrovertible, by the loudest and most charismatic mouths.

I propose to follow Dewey’s notion of truth as warranted assertibilility, with warrants of logic, facts and practical workability. Now, warrants of logic and facts are often ignored, and are replaced by emotion-laden appeals ‘Is’ is replaced by ‘ought’. I allow for emotions, value them, but this killing of rationality spells disaster.

Science is held to yield the progress of truth, but has a considerably conservative practice. As discussed before, according to Imre Lakatos (1978) science takes the form of research programmes, collections of theories that have a shared hard core of basic assumptions and methodological principles, and a surrounding protective belt of subsidiary assumptions. When a theory in the programme is falsified, the core is left unaffected, and adjustments are made in the protective belt. The economist’s drive for calculation of an optimum belongs to the core of the discipline of economics. When a theory is falsified, the principle of maximisation is not challenged, but the function to be maximised is tinkered with. I once received the following rejection of a paper submitted to an economic journal: ‘this is not optimisation of an objective function, therefore it is not science’.


-          Is there any certainty

-          Is there fixed truth

-          What certainty does science give

-          Is evolutionary theory a good approach in the humanities

-          What is the rationality of theoretical conservatism

-          What scenarios would you propose to investigate


No comments:

Post a Comment