If you are a founder, the first question you need to answer is “What problem does your solution solve?” Unless you can answer this question succinctly and accurately your project is likely to fail. And yet founders often find this is the hardest question to answer.
The next question- understandably- asks “Is your problem worth solving?”. This forces you into the murky territory of judgement. Mostly, this is codified in terms of market size and potential return. The third question is who is already solving this problem? And the fourth should be what are the second and third order effects of solving the problem? We don’t do this too often, but so far, so systems thinking
But it can also be important to recognise that many of the ‘experts’ who will advise you in answering these questions also -despite their qualifications- may not have answered these questions in their own disciplines. And yet, their work and input may support your new AI radiographic diagnostic application or your workplace emotional monitoring app. However you dress up your AI, if the “experts” in your expert systems aren’t “really” experts you will face a problem at some point. Sometimes an irrecoverable one.
And this is where relabelling becomes important. You can- if you wish- relabel any product. And almost any academic or professional discipline, if you have a mind to. This is almost normal once new tools such as computation are introduced. Statistics- dry, dusty, boring, incomprehensible- becomes data science- sexy, fluid, pragmatic and fun. “Look at this visualisation!”, “See the correlation between the plum harvest and the new strain of COVID19. All you have to do is chop down the plum trees! Or plant more!”.
Because every discipline has a history marked by stable epochs of confidence and theoretical coherence that alternate with times of uncertainty and conceptual confusion. These things are related. If you have a theory and it works you feel good. Argument is limited because the existence of an accepted theory calms things down a lot. The sun goes round the Earth, right?
And, of course, scientists being human, tend to follow their existing intellectual, temporal and financial investments- as do we all.
The relatively new field of psychology has passed through several such phases since its genesis in German universities in the late nineteenth century. This migratory path was initiated by a hybrid of philosophers and physiologists who believed that human perception, thought, emotion, and morality would be illuminated by studying the relations between externally measured biological processes and the self-reported or observed felt mental processes that distinguished humans from other animals.
Processes such as thinking, (cognition) feeling (affect) wanting (volition) and operating socially (interacting) at a range of different levels.
But if you are going to make claims for a new product- particularly one which claims evaluative, descriptive and predictive power you need tools and blueprints which are unassailable. And- handily- there was already a natural science which had proved to be extraordinarily explanatorily powerful. This is why psychologists chose physics as a model for their tools, concepts and metrics. The ability to separate out the objects and concepts in the experiment proved irresistible.
And when computational data gathering methods were added to this, the idea of behavioural science became firmly embedded in advertising, marketing, and many other disciplines.
But these processes are part of an ecosystem such as that we find in biology and population ecology. Had they chosen these disciplines, they would have been more concerned with the processes that transform data into knowledge and the thinking, feeling, wanting processing states, as patterns of evidence.
This would have helped them recognize that the specific setting in which measurements were recorded often limited the generalisability of the abstract
concepts they used. And, of course Unfortunately, these first investigators were hampered by a lack of methods to measure brain activity in living humans. This meant that they focused on the nature of external experience based on their own interpretations of events.
Of course, cognition, affect, volition, and interaction (CAVI) — together with their “intended” and “unintended” consequences are far more complicated than the ability to detect a difference in the loudness of two noises or the rate of decay in a set of recently memorized syllables. This led, as it often does to a practice schism
In the United States psychologists, often more pragmatic and egalitarian than Europeans, addressed the study of how behaviour arises in animals, because these processes are easier and perceived as more moral to measure. One advantage of this research was that it gave psychologists a better understanding of the robust mechanism of “conditioning” as a way of explaining, predicting and controlling some changes in behaviour.
This approach produced a fair share of victories during the half-century when behaviourism, led by individuals such as Skinner, Spence, Miller, and their students dominated many psychology laboratories. Behaviourism promised to explain, evaluate and — in part- predict the emergence of a set of universal behavioural properties in each species as well as the variation among individuals within a species as the result of differential opportunities to develop conditioned habits. Cognition, affect, volition and interaction, however, remained resistant to explanations limited to conditioning principles.
In Europe, research focused on the mind as accessed through both self-reported states and observed experiences. This led to a growth of interest in psychotherapeutic explanations. Adler, Jung, Freud and others developed
a series of initially persuasive set of explanations that, seemed intuitively correct to many. In the case of Freud this was because they supported two popular beliefs. The first, that the events of early childhood placed serious limitations on a child’s development and the second that reproduction and sexuality played a significant role in human interaction. Obvious, no?
Psychologists and users of psychological theories nurtured the hope that the experimental success of conditioning and the unprovable concepts of psychoanalytic theory could somehow be integrated to form a unified
theory such as that developed in quantum physics — a theory of everything (TOE) which would- finally- explain behaviour and find a way in which it could be described, analysed, evaluated, explained, predicted and -self or other-controlled.
This hope resurfaced with the development of machines which enabled us to ‘track’ electrical energy, chemical presence/absence and blood flow volume and rate in the brain. Such technologies included positron emission tomography (PET), electroencephalography (EEG), magnetoencephalography (MEG), and different types of functional magnetic resonance imaging (fMRI), to determine the physical origins of psychological events.
This type of research promised to provide psychologists with a set of fundamental structures that would be analogous to particles in physics and cells in biology.
By looking at the interactions within and between such structures using the technologies identified above, it was hoped that we could map behaviour and, indeed, mental ‘health’ and find approaches that helped ‘manage’ both.
This has been fruitful in determining the potential functions of particular physical ‘brain’ networks such as the “executive functions” of the prefrontal cortex or the ‘salience network” which directs attention. But is this more than the Cartesian pineal gland which was advanced as the ‘seat’ of perception in the 17th Century?
This research rests on the premise that specific patterns of activity in particular neural ‘networks’ will ultimately explain why a person behaves in a particular way. By reading these ‘signals’ we can build predictive models such as those used to identify the onset of an epileptic episode.
Of course, there is increasing evidence that some of these activity patterns may be initiated by signals from the gut biome. And indeed there are models which suggest that
Another approach, was to study factors such as intelligence, personality traits and/or mental illness, and to seek patterns of behaviour at one point in time or over time which could be tracked onto events. Unfortunately, these psychologists did not possess a set of powerful methods to support their hypotheses, and many settled for the traditional strategy of actually asking people what they felt, believed, or did.
Over the past half century, psychologists have administered a large number of questionnaires with unique names to both adults and children that were measuring similar phenomena. Sadly, language provides limited information on most mental states because all people have a restricted vocabulary, are tempted to conceal traits they regard as undesirable, and have limited access to what their brain or body
is doing when they describe their feelings or motivations. Ninety percent of the brain activity which ‘initiates and maintains’ those actions is unconscious.
So, are we trapped in a cascade of incompetence as the gap between our physical body and our minds remains uncrossable for consciousness? Well, apparently not.
There seems to be a phenomenon which provably exists but which has also been called a ‘many-headed monster’. This is metacognition. The simplest definition of metacognition is cognition about cognition- thinking about thinking. A reflexive
metacognitive process is meta-level with respect to an object-level cognitive
process. Mindfulness might be seen as an example of metacognition. So might the feeling of a tip-of-the-tongue experience
Importantly however, these examples of metacognition are associated with explicit conscious awareness. While metacognition may be accompanied by conscious awareness in humans, this need not be the case, suggesting a division between ‘‘explicit’’, conscious metacognition and ‘‘implicit’’ metacognition.
Tulving argued that there was a traditional bias in psychology to assume a strong correlation between cognitive processes, behaviour, and experiences. This is, given the nature of the discipline and its’ lay applications that a particular cognitive process, such as retrieval, is associated with a particular behavior, a verbal description of an earlier episode, and that this behaviour is always associated with a particular conscious experience, such as “mental time travel”- recalling, say, a childhood home which no longer exists.
Tulving further claimed that this model no longer worked — there were too many demonstrations of conscious experience not accompanying a particular behaviour to warrant its challenge. He cited studies on implicit memory, in particular, in which memory processes create a change in behavior, but without the accompanying mental experience. More recently, we can point to research in which mental experiences of memory arise from cognitive processes not tied to the retrieval process. For example, Cleary and his colleagues showed that déjà vu experiences arise when a familiarity experience occurs without corresponding retrieval of event details.
So, it seems possible that we can increase awareness and control of unconscious processes. In the next paper we will look at how this might be achieved.
And how it might impact on the effectiveness of founders, and the rest of humanity.
Oh yes. And break the cascade of incompetence.
Thank you for your attention.