Question the Quality of our Categories

Things and Groups of Things

Our brain has evolved to perceive the world as a collection of things -

• individual things, objects, and
• groups of things = categories = classes = entities = types.

It makes sense to group things together into classes /categories - IF they are usefully similar;

• in their nature,
• in their behaviour,
• in the given context,

because - IF we know and understand a thing's type, its nature – then we know what to expect - in terms of its;

• essential properties,
• optional properties,
• behavioural repertoire and triggers,
• and the possibilities and probabilities of its life-cycle.

So when you encounter a group / category / class that;

• someone else has composed, or
• your culture is presenting as a default category (probably via the media),
• ask -
• What are this category's membership criteria?
• What are the essential and optional properties a thing should have, in order to be considered a member of this grouping?

Things have properties = attributes, features.

There are 4 important sorts of properties;

• unique identifiers – ways to identify a specific member of a group,
• essential properties,
• optional properties,
• recurring properties – lists, histories, e.g., previous addresses.

Properties have names and values.

The values may be restricted;

• one only – as in a birthday,
• select one from a limited check list – animal or vegetable or mineral,
• select any from a limited check list,
• free within possible range limits – as in age,
• (may be a smooth or lumpy continuum),
• multiple current values,
• e.g., people may have more than one current email address, vehicle, job,
• multiple historical values.

Things have life cycles – cycles of probability and possibility,

• an acorn may grow into an oak tree – but not into a willow tree.

Things can have a repertoire of many behaviours (called methods in computer speak);

• behaviours affecting themselves,
• behaviours involving or affecting other things,
• behaviours can be context sensitive, triggered by particular circumstances,
• behaviours can change / evolve over time,
• history of previous behaviours,
• current behavioural repertoire (possibilities),
• common habits (probabilities).

The quality of our categorical groupings is very important but sadly nebulous – because our culture does not yet have commonly understood words to describe the quality of our categories. We need to invent some catchy category-quality describing words - to help us reflect on, and assess, the quality of our categories.

There are two basic quality issues –

(1) the membership criteria and the resulting degree of similarity,

and

(2) are the criteria commonly understood?

1) The Basis of Group Membership

and the degree of similarity it produces.

ASK yourself - is this group’s membership based on;

• an appropriately large number of essential attributes (e.g., a traded commodity, a particular model of washing machine), or,
• a misleadingly small number of optional attributes (e.g., supporters of a particular football team)?
• Knowing it’s a Bosch Exxcel 1400 is the key to finding out everything you are likely to need to know about a particular washing machine.
• Knowing that someone is an Exeter City supporter tells you nothing about their; occupation, accommodation, education, nationality.
• Knowing that an animal is a Hereford cow probably tells you a lot about it.
• Knowing the weight and quality of a gold bar, in troy ounces and carats, certainly tells you all you need to know about it.

So – to repeat – ask yourself - is this category's group-membership based on all essential properties or just a few optional properties? It's a sliding scale.

The Equality Dimension

the second aspect of the membership quality issue - can be highlighted by asking - are all members of the group equally good members of the group – or do they vary = are there good and bad examples of this category?

• The types of group/sets we use in logical thinking usually assume that all members of the set are equally good members of the set (prime numbers, all men are animals).
• But the kind of categories that our neural networks create can be much more subtle than that. For example, we all intuitively know that robins and ducks are members of the bird class, but cross-cultural studies have shown that we consider robins to be better birds than ducks – ducks swim under water, can eat under water – they are quite fishy in some ways, and they don’t sing, they quack. Ducks are birds, but robins are better birds than ducks.
• The classes, Robins, and Ducks, are both; more consistent, more similar, more specific, than the more general class, Birds.

That is why general classes are prone to having more and less representative members – which could/should probably be grouped into more specific sub-classes – if it is useful to do so.

We use a lot of these general classes - but we hardly ever stop to ask if John is a good or bad example of a - Racist / Fascist / Socialist / Banker / Refugee.

This is a huge weakness in our human thinking.

We are capable of making much much much finer distinctions.

Practise challenging over-general categories, and making fine distinctions - where you can, if it is safe to do so, if it matters.

Encourage other people to question whether X is a good, or a bad, or a perfect, or an imperfect example, of Yness?

Sometimes we enshrine over-general categories in law. Particularly laws which seek to implement universal principles or rights.

We seem to feel that we should not make subtle distinctions or set any limits on things we have defined as 'universals'.

But what if it was a mistake to call them universal - what if the reality is that 'universals' get a bit fuzzy and relative and complex and contradictory near the edges - when you look at the whole dynamic interconnected big picture.

• So, for example, we have laws preventing discrimination on the grounds of religion - but religion, and discrimination are both hugely general categories.
• What if a new religion required, or allowed, or encouraged its followers to do things which were already illegal in that society? The anti-discrimination law would allow religion to be used as a trump card for breaking those pre-existing secular laws.
• Surely it would be more coherent to use more specific categories, and define what types of discrimination are, and are not, allowed, in relation to specific, illegal but religion-justified behaviours and beliefs, and in which specific contexts these secular behavioural requirements and exemptions apply.
• We have done something similar with the over-general concept of asylum, creating a highly desirable trump card which appears to override all other valid national entry requirements. This trump card is so desirable and valuable that an international network/industry has sprung up to share in that trump card's unique value. The asylum definition could easily be adjusted in the light of experience, so as to better achieve the original humanitarian intention, whilst reducing its exceptional trump card value.
• There is an asymmetry built into the system, in that the people smugglers (individuals and networks) are inter-nationally distributed and beyond the legal reach of the destination countries that bear the costs and responsibilities of providing asylum. Surely the conventions and legislation (the DNA) that created this system should be adjusted in the light of experience to address this asymmetry, and to balance the powerful and international temptation to abuse this very special privilege, with a matching set of very special and internationally effective punishments for all those participating in the deliberate abuse of this privilege.

2) Is It Universally Understood?

The second Category Quality Issue. Is the class/category’s meaning universally understood? Are the membership criteria, and the degree of similarity, commonly understood by the 'stakeholders'? Usually not!

• A good category tells you a lot about each of its members – a bad category appears to - but doesn't.
• It's all about the precision with which you can reasonably infer knowledge about a thing, based on its suggested membership of a suggested category/class. How much of its identity is dependent on, determined by, can be inferred from, its suggested class membership - and to what extent is its identity independent of the suggested class membership.
• So - in order for classification/categorization to be a useful communication and thinking tool it is necessary for all the parties to the communication to have a common understanding of how the class has been constructed – (the extent and the proportion of the essential and optional membership criteria).

Usually we don’t have that shared understanding.

Sometimes the lack of clarity is accidental, but often it is deliberate – because the powerful emotional impact of positive or negative association (in the fast-system) is immediately broken if we engage the slow-system and start trying to precisely define the essential and optional properties required to be considered a proper x, y, z.

• Look at the daily news – it is littered with poor quality, undefined, un-agreed categories, being used to trigger emotional associations. Our media continue to do this, because we continue to accept these emotional associations without question or complaint.
• If we challenged the quality of the media’s presented categories – then their attempted emotional hijackings and cognitive manipulations would be exposed for what they are, and the perpetrators would be discredited.

The fundamental reasons for evaluating and challenging the quality of presented categories, are

• the pursuit of truth, and
• respect for reality.

There are also many reasons why we don't challenge bad categorization - why human groups have taboos, fudge issues, and ignore elephants in the room. These reality distorting strategies are rooted in the evolutionary advantage enjoyed by coherent, vector-aligned groups, over diverse-vector (but truth directed) groups.

So if you are interested in the pursuit of truth and respect for reality, you should be asking;

1) Quality - where does this proposed category’s membership criteria sit on the sliding scale - all essential properties to one optional property?

2) Consistency - are all members equally good members of the group? Are the members identical or consistently similar – or is there a significant degree of variety?

3) Context - what membership criteria/properties do we need to be concerned with in this specific context? Can we accept the default classifications or do we need to redo our things-groupings - for this specific context.

4) Agreement - are the membership criteria and degree of similarity clearly understood by all the 'stakeholders'? If not, there will be complex misunderstanding and confusion.

So - categories vary in quality. They also evolve - they adapt to changing circumstances and contexts.

When expressed in language, they are inevitably less subtle than the categories represented in our neural networks – because our personal neural networks can make finer, more numerous and more subtle distinctions than our shared vocabulary can describe.

My politics is far more subtle than left or right.

My judgements about people are based on a huge number of subtle distinctions, not just on skin colour, country or culture of birth.

Good and Bad Categories

In the best categories - all members are functionally identical and there is total clarity, understood by all concerned, as to the membership criteria.

• Examples – traded commodities – a troy ounce of 22 carat gold, a barrel of Brent Crude oil.
• What shall we call them? – hard categories, precise categories, complete categories, perfect categories, valid categories.
• Consider. Discuss. If you come up with something catchy – spread the word.

In useful everyday categories – the members are not identical, but are consistently similar in all important ways, in relation to the relevant context, and the membership criteria are commonly understood by all concerned.

In over-generalized categories – the real world distinctions and similarities are much more subtle than our simplistic general naming system can convey – and thus, members are not all equally good members of the group. Birds; robins and ducks share many essential attributes, but some aspects of their lifestyles are quite different – but in daily life it is, on balance, an aid to communication.

• If you are standing on a river harbour wall and point in a particular direction at a scene including a duck and a boat, and say, “Look at that bird,” people will focus on the duck and not on the boat. (Of course – boat is another very general category.)
• What shall we call these over-generalized, simplistic, quick but vague, do for now but no good for detailed thinking, categories?

Public relations and propaganda agents often use emotional-association categories – designed to invoke emotions of; like or dislike, good or bad – triggered by a suggested link or association with an existing, simplistically named, good or bad, category, whose membership criteria are either; not specified at all, or not in sufficient detail, and therefore, are not commonly accurately understood.

• This car is sexy. UKIPers are racists. Tories are selfish. Public sector workers are good and need your money to help meet the 'needs' of the 'vulnerable'. Pointing out flaws in NHS performance shows a lack of respect for this great institution which suggests your real intention is to destroy it……
• We could call these deliberate (or accidental), emotionally manipulative propaganda categories. Can you think of a name that encapsulates these key aspects in a snappy phrase? Consider. Discuss.

Grouping things on the basis of one optional attribute is a really bad use of categorization. For example; people who happen to support a particular football team. This is like making up a group name for the things that just happen to be in a junk shop or a scrap yard.

• The fact that they are in the junk shop, is all they definitely have in common.
• The only thing that Exeter City supporters definitely have in common – is that we call them Exeter City supporters.
• We could call these ‘junk shop,’ or ‘scrap yard,‘ or ‘name only’ categories – because there is just a name, based on a single criteria with no attempt at any further membership criteria definition.
• Of course, if the context is, deciding which entrance to use at the football stadium, then it is a very appropriate classification. But if you are being encouraged to like or dislike someone because there are a 'lefty' or a UKIPer, or a 'banker', or a 'public sector worker', etc., then that is very shoddy thinking.

Learn to recognize these ‘name only’ categories – question whether or not they carry any useful meaning or reliably imply any other (dependent) attributes. Challenge them where it is necessary and safe to do so.

Today - a national radio station, which to its credit, dedicates each hour to hearing callers' views on a particular topical issue, asked whether 'immigration' had been a good or a bad thing, and whether 'public sector cuts' were a good or a bad thing.

• 'Immigration' and 'public sector cuts' are both hugely-general name-only categories that are crying out for finer distinctions to be made. But there is an enormous social psychological cognitive power in the public over-general framing of these questions because the issues have been politicised - turned into groupthink signifiers. Not one of the callers suggested making finer distinctions. All accepted the 'either good or bad' framing.
• The category 'immigration' includes millions of unique individuals and all the consequences that have followed from their migration and cultural dislocation, for everyone affected.
• It is a bonkers category, crying out for more subtle distinctions. And yet we fall for this simplistic emotional triggering - over and over again.
• 'Public sector cuts' - well cutting down on; waste, inefficiency, poorly defined counter productive goals and perverse incentives, breaking-up fraudulent self-serving vicious circles and networks that siphon off money that should have been used to benefit the public, exposing and punishing corruption and conspiracies to cover-up failure and abuse - all sound good to me, but not perhaps to those who benefit from, or are trapped in, such activities. But education (knowledge and thinking), skills training, re-training, health promotion, health care, proportional policing and genuine resource-pooling risk-sharing safety nets, all sound like good things that should not be cut. So to my mind it is another bonkers category, crying out for more subtle distinctions.
• Opinion polls often ask 'this or that' questions about hugely general junk shop categories - "Do you feel the Government is doing a good job - yes or no”?

Look out for bonkers, emotive, groupthink-signifying categories crying out for more subtle distinctions.

One or Some equals ALL

Perhaps the most serious mistake we make when using categories, and one that we make over and over again, even though we know better, is making careless and flawed assumptions as to what you can reasonably infer from membership of generalized categories.

• When thing X has one, or some, of the many essential and optional attributes of class Y – we somehow assume that X has all of the essential attributes of class Y – wrong, wrong, wrong.

This is a huge impediment to human debate and group cognitive evolution – so I will say it again – because of its associative nature, the human brain is very easily tricked into wrongly assuming / concluding that because X has one attribute, that is one of the many optional or essential attributes of class Y - it is therefore a full-blown highly representative member of class Y.

This one=some=all trait, is an inevitable consequence of the way our neural networks detect associations. Detecting associations has been a huge benefit;

• it is an essential element of our amazing perception and meaning-making abilities, detecting objects, recognizing patterns, forming rapid assessments of complex situations despite noisy and inadequate data,
• and it embodies the precautionary principle - assuming that glimpse of colour could be a tiger.

But this trait has a down side. It can, and frequently does, trigger some very rapid jumping to some very wrong but highly believable assumptions.

If you think about it for a day or so, you will notice that we make very rapid decisions about people and situations, which are not logically justified by the available information. In a few seconds our senses notice flags, triggers, signifiers, that fire off (by neural association) large packages of assumptions - popularly called stereotypes -

• these cognitive leaps can be triggered by; appearance, posture, body language, vocabulary, accent, grammar, idioms, dress sense, hair style, body modifications, employment history, tomato sauce on chips, preferred drink; beer or lager, g & t, wine or spritzer, hot chocolate or coffee - choice of car; off-road adapted 4 by 4, 2 door jaguar, eco dual-fuel-hybrid - types or reading material, newspapers, favourite TV programs, leisure activities; football supporter, rugby supporter, show jumping fan, golf enthusiast, cinema, opera, theatre, religion and sect, political allegiance, type of school, quality of university, choice of subject specialisation, level of education, type of qualifications.

The impact of this down-side (the logical one=some=all flaw in this otherwise very valuable thinking ability) can become greatly exaggerated if it gets entangled with groupthink and group-membership dynamics - particularly if there is what might be called a group orthodoxy involved - which is very common;

• approaches to education, reasons for education, causes of climate change, nature vs. nurture, what constitutes social progress; subsidize or empower, protection or evolution, borrow or earn.

Powerful forces (reward and punishment) encourage group members to adjust their view of the world to fit in with the group's orthodoxy. This strengthens group boundaries, filters group membership, and aligns group vectors, but this group dynamic is no respecter of truth or reality. It distorts our perception of reality. It is not a truth-directed process.

Emotive Propaganda

This kind of logical error (One=Some=All) can easily happen accidentally, but often, it is a deliberate act of emotional propaganda, an emotional hijack, an emotional racquet (look at me – I am good – you are bad, or worse still, look at us, we are good - you are bad) - in which case it usually comes hand-in-hand with an aversion to defining precisely what they mean by Y - an aversion to defining Y's essential and optional attributes - because the perpetrators know, intuitively, that if we engage our slow thinking system, it will break the power of the emotional association.

We could call this an emotional hijack by trivial association.

• It is an emotional name calling exercise (left wing, right wing, separatists, freedom fighter, terrorist, insurgent, pro-R-rebels, moderates, extremists), with no attempt to clearly identify and agree the category’s membership criteria in terms of its essential and optional attributes and behaviours.
• It is an attempt to stop any deeper analysis or more nuanced thinking - and it often works.
• It is a racquet. It says, "For good people, like us, this simplistic association is all the thinking you need to do, in order to like or dislike this thing, person, idea."

When the selection criteria that define a particular category are not generally understood and agreed – then references to that category will probably get in the way of cooperative, collaborative, constructive thinking.

It will be a major issue, and a major cause of confusion in any attempt at cross cultural, cross system, cross boundary cooperative modelling of a situation.

Recognizing accidentally flawed and deliberately manipulative categories in commercial and political propaganda is an important skill to develop, and a major component of effective systems thinking – because flawed categories drag you away from reality.

Finding a constructive way of reacting to deliberately flawed and manipulative categories is the holy grail of diplomacy. Putin's and Lavrov's patient considered responses, to the West's recent campaign of manipulative emotive categorical propaganda against ‘Moscow’ (a city?), has been a master-class of dignified cognitive self-control, that we may be able to learn from one day - but not yet. The groupthink is far too strong at the moment.

Finally – detecting and grouping things, is a core neural network ability.

The human brain is capable of, and usually is, operating very subtle distinctions - much more subtle than either the available vocabulary, or the slow-system's logic, can describe.

The down-side is that it makes us prone to association errors, which can be exaggerated by the dynamics of group orthodoxy.

I-think usually has more subtle and more easily up-datable categories than groupthink. Consider. Discuss.