Dialects of Reason

Oli Sharpe, August 2013


 Download a PDF version of this essay

Conversation is one of the great pleasures of life. Sitting in a café with friends debating the issues of the day is an enjoyable way to spend time together, but when we disagree what exactly is going on? How can we know which arguments are sensible, well reasoned arguments and which are just rhetorical attempts at persuasion? Indeed, is there even a difference between the two? This is a brief essay (or long blog post!) that expresses my current thinking around what it means to me for an argument to be well reasoned.

Modernism, post-modernism and spin

We live in an interesting stage of history in relation to the idea of ‘reason’. It could be said that the early years of the 20^th century were when reason was held in its highest regard. Since the Enlightenment the awe inspiring abilities of science and engineering to increasingly explain and control the deterministic world around us have been helping to replace superstition with reason. This trend was philosophically supported by successes made in the early 20^th century towards underpinning all mathematics and science with a formal logical framework, such as described in Principia Mathematica. Logical reason was helping build a better, new world and a modernist, rational utopia beckoned.

But by the middle of the 20^th century the modernist agenda had been dealt a series of self-referential setbacks. Gödel’s incompleteness theorem used logic to show that no axiomatic system could capture all of what we consider to be mathematically true. Philosophers wrote texts that revealed the deep problems that occur when we as members of a given society try to use language and science as tools to describe the world objectively. Writers from the elite laughed at figures of authority when viewed from behind the wizard’s curtain. With quantum mechanics and chaos theory fundamental uncertainty became core parts of the dialog of science and mathematics. Politically the terrible crimes of the fascist nationalists could be seen as an extreme and terrifying interpretation of the modernist agenda.

And so with all these new concerns about the limits of reason, post-modernism, or whatever you want to call it in your text, rationally undermined the dream of a rational utopia. The link between formal logic and reason had been weakened.

So where does that leave us at the start of the 21^st century? Now we are in an age where politicians speak about ‘creating reality’ in order to further their agendas and the science of climate change is dismissed in some popular media as being part of a left wing conspiracy rather than having its complexities debated intelligently. We live in an anti-intellectual, narcissistic society where everyone wants to be famous and too few people want to study science and engineering. Experts are viewed with suspicion whereas, despite the global banking failures, the market is revered as a magical, neutral solution to anything that can be traded. And rather than learning from the terrifying consequences of extreme nationalism last century, far too many politicians are dangerously awakening the daemon of tribalism as a distraction from the genuinely tough economic problems faced by so many countries.

One interpretation of the state of society today is that post-modernist concerns have led to an ‘anything goes’ society where the value of rigorous intellectual effort is questioned or even treated with suspicion compared to the hearty truths that come from the gut of everyman. The so called ‘information age’ hasn’t propelled us into an age of ever deeper reason but rather into an instantaneous age of celebrity, corporatism and political spin.

This seems the appropriate place to bring myself into the narrative. Over the last few years I have been struggling to reconcile two seemingly opposing opinions of mine. The first is a deep seated belief that, despite the concerns of post-modernism, our ability to reason is key to solving many of our most pressing global problems. We need reason. The second is the sense that we all regularly express opinions and behave in ways that reveal poor reasoning. Watching the daily news it is easy to fear that there is a profound lack of good reasoning in the world. And reflecting on my own thoughts and arguments it is often not clear to me that the positions I hold are supported by a rigorous level of deep reasoning that I could honestly defend in depth.

For example, imagine the subject under debate is whether or not nuclear power is safe enough to use. From a modernist perspective this should be precisely the kind of subject where evidence based reasoning delivers a clear cut, rational answer that everyone agrees with. However, it is often the case that two well educated, intelligent people have completely opposing views on this subject.

In any given dialogue between people there are obviously other things going on, such as the attitude that the interlocutors are taking to each other, but in this essay I am only concerned with understanding dialogue between people who are being genuine, honest and considerate towards each other. Even with the best attitude, there are still many vigorous disagreements between well meaning people where it would be good to understand what is going on.

For example, I think we need to use nuclear power as part of our portfolio of energy sources, but I know many other sensible people believe that nuclear power is just too dangerous to use at all and I can imagine having an intense debate over just this kind of subject.

Most such debates happen between individuals who are lay people on the scientific details at hand and yet the debaters are often intensely confident in their position, sometimes to the point of considering the opposing view to be very poorly reasoned and even ethically questionable. Indeed, most of us would like to believe that our opinion on such subjects represents a well reasoned position, or even the rational position, when in fact we’re mostly repeating some mishmash of received opinions that we first read or heard in the media or some other source.

How does someone know that nuclear power stations are too dangerous? Because they’ve read it. Why do I think that nuclear power can be safe enough to use? Because I’ve read it. So if we sit around debating this point are we really just saying: “The expert I trust disagrees with the expert you trust”? Is this really reasoned debate or simply rhetorical posturing?

This is just one sketch example of how the supposedly solid foundations of reason can slip away the more closely you examine what is going on. Indeed, as a post-modernist would argue, it appears to be impossible to provide an objective dividing line between spoken arguments that would be universally recognised as well reasoned as opposed to just being rhetoric. So, what then is reason?

What is reason?

Maybe the best way to look again at reason is to start from how we use reason on a daily basis.

We give reasons for our past actions and offer reasons why someone should follow a given course of action. We judge others as having ‘good reasons’ or ‘bad reasons’ or even ‘no reason’ for their actions or opinions. We explain to others the reasons why certain things happen, or indeed what we understand to be the reasons why someone did what they did.

In general, it could be said that ‘reasons’ are what we try to respond with when asked a ‘why’ question. ‘Reason’ is then simply the capacity to provide such answers. Someone is ‘reasoning’ when they are in the process of constructing such an answer, such as constructing an argument. Notice that using reason to decide ‘how’ to get something done could be seen as preparing to be able to answer ‘why’ did you do it that way.

The interesting question though, is how do we decide which answers are ‘well reasoned’? How do we distinguish between good quality reasons and poor quality reasons? What is a good argument?

A classic answer to this would be that a good argument is one that is logically sound (in other words the structure of the argument follows formally correct logic and the assumptions of the argument are true). This is where my views on reason differ from many others.

Rather than seeing reason as our attempt to be logical, I see logic as our simplistic attempt to model reason. I say simplistic because it fails to capture what is going on when we reason more often than it succeeds. Not only is formal logic subject to all the concerns of post-modernism but also the formal notion of a sound argument clearly fails to coincide with the day to day notion of a ‘good argument’ that most people around the world and throughout history would recognise.

Indeed, if reason were our attempt to be logical then humans fail regularly (indeed sometimes appear to fail willingly) and therefore we could not see ourselves as creatures of reason. Rather reason would be something that only the most Vulcan of us achieve in the rarest of circumstances. And yet in common parlance we do all regularly use reason and provide good arguments that our peers would agree are well reasoned.

So, if we are not trying to be logical when we reason, what are we doing? And how can we judge an argument to be well reasoned or not if we can’t appeal to the notion of a logically sound argument?

A subjective notion of reason

Maybe the problems with a logical notion of reason arise because it’s an attempt to be an objective notion of reason. What if we start from what appears to me to be a well reasoned argument?

I seem to have an internal ‘model’ of the world which I use to decide which actions to take and how best to respond to events in my environment. Obviously I don’t mean ‘model’ in the sense of an internal 3D model of the world, but in the sense of an implied, informal theory of the world that I use to try to comprehend and predict the world around me and how people and objects in the world would behave in response to various actions I could take. If I throw a ball I know roughly where it will go. If I were to snatch a newspaper away from someone reading it I know they wouldn't be happy. We all use our internal model of the world to operate effectively in the world. This model is not something separate from us it is a big part of who we are.

Now, if in a debate someone makes an argument that seems to be plausible with respect to my model of the world then it sounds reasonable to me. If, however, someone suggests an idea that doesn't fit at all with my model of the world then it sounds wrong to me. It sounds poorly reasoned. Maybe reason is as simple as that.

Furthermore, if an argument fits poorly with my model of the subject at hand, but fits well with my model of what a good argument sounds like in general, then I will see the argument as being good rhetoric rather than good reasoning. In this way the distinction between rhetoric and reason is a completely subjective distinction.

I have been referring to this model in the singular and I will continue do to so, but it is worth noting that it is not a single well integrated, coherent model of the world, but rather it is a mishmash of mini-models that are used in this context or that. Often these mini-models are based on how we've seen other people behave or argue in this context or that. There are some contexts over which we have reflected deeply and therefore we have a complex mini-model for that aspect of life. However, my experience of myself and of most people is that there are many areas of our lives for which the mini-models with which we operate are actually quite simple and unexamined mini-models that we have largely learned from others.

The fact that we all operate with such a mishmash of potentially inconsistent mini-models in different contexts is a key reason why we are all likely to hold contradictory views if someone were to try to map out all our beliefs into a formal logic. So even our own set of beliefs and values would fail to be logically sound in a more traditional view of ‘reason as logic’.

In contrast the general proposal here is that something seems well reasoned to a particular person if it coincides sufficiently with their internal model of the world for the given context. Conversely, our ability to reason is our ability to reflect on and express to others aspects of our model of the world which they then will judge against their own model.

We often explain our reasons poorly, or even get them slightly wrong in our first attempt to express them. To me this shows that our attempt to explain our reasons comes after the work done by our model itself. So, reason is not our ability to model the world, but our ability to engage in dialog about our models of the world.

At first glance this definition of reason, with no explicit notion of ‘Truth’ or ‘logic’, appears to be very weak and at risk of allowing unlimited relativism, but this is not so. The crucial constraint on reason within this definition comes from the link to the world through the individual’s attempt to develop a useful model of the world.

If an individual were to have a completely useless model of the world with no predictive capability then this would render them quite incapable of operating in the world. Typically people living such a chaotic, out of control lifestyle appear to others to be behaving without reason.

Conversely, the more effective an individual’s model is, the more likely they are to be able to operate successfully in the world. As science has discovered, mathematics and logic are useful tools when building predictive models of the world, and so too these consistencies are likely to be important aspects of our internal models (whether explicitly or implicitly) and thereby most of us have a strong connection between an argument seeming logical to us and it being well reasoned. Indeed, for me the word ‘rational’ relates to reasons that rely more heavily and explicitly on logical arguments.

Local dialects of reason

The idea that we unconsciously judge something someone says as sounding well reasoned to us by ‘testing’ if it fits comfortably enough with our internal model is similar to the way that we unconsciously ‘test’ whether the structure of their spoken sentences sound grammatically correct to us or not. Indeed, one could imagine these two functions as having a shared origin or even one having a dependency on the other.

The important point of this analogy with grammar is that however strongly certain linguists consider their grammatical model of a given language to be ‘correct’ there will be many speakers of that language who would either disagree, or at least would not speak in a way that the linguist considers ‘correct’ but these speakers would be perfectly well understood in their own community.

So the proposal here would suggest that there will be many ‘local dialects’ of reason within which the people consider other members of their community to use better reasoning than those outside of their community.

This is why there are so many people in the world who appear to me to be behaving in ways that make little sense, while they would insist that they have good reasons for their approach to life. They are simply using a different dialect of reason, which they can use very well but I can’t. We’re all thinking and acting with good reasons with respect to our internal model of the world, it’s just that in some contexts we have some major differences between our models.

However, given that we all share the same world many aspects of each of our predictive models of the world are likely to have similarities even between disparate communities. Hence our sense of what is well reasoned will often coincide sufficiently to give us an impression of an inter-subjective rationality that on a day to day basis can often feel like we’re all working with the same objectively ‘true’ notion of logical reason. Again, using the analogy with grammar, when it comes to reason we all appear to be speaking local dialects of roughly the same language which is what allows us to have at least a basic dialog with anyone.

Indeed the project of science could be seen as a project to build an explicit, external and shared model of the world against which we can all develop an increasingly advanced dialect of reason for certain subjects. The success of this shared model can make it easy to fall into the illusion that this model is the objectively true model that can form the basis of an objectively true notion of reason. However, not only is the methodology of science confined to certain types of subjects, but also post-modern concerns will always render the outputs of science just short of being truly objective.

Some immediate consequences to note

There are some immediate consequences to this view of reason. Firstly, it means that it is impossible for a subject to hold genuine opinions about the world that they themselves consider to be badly reasoned. This is because to be a genuine opinion it must be consistent with their world model and therefore would appear to be well reasoned to them. Everyone really does think that they are right.

Secondly, while colloquially many people tightly associate a well reasoned argument with being logically consistent, here we are instead saying that there is no such requirement. Rather, someone is acting with good reason if they are acting in genuine accordance with their model of the world, however logically inconsistent that model may appear from another perspective.  Despite these inconsistencies, we typically suffer from the illusion that our thoughts and actions are logical precisely because to us they seem well reasoned and we associate reason with being logical.

Thirdly, our ability to predict the likely consequences of our actions comes before our ability to judge whether we think a particular argument is well reasoned or not. To me this implies that a consequentialist grounding of ethics that uses this predictive model to understand the impact of our actions on others comes before we can refine this further with a logical exploration of ethics or a mathematical utilitarian calculus. This is obviously a bold statement that I hope to explore in another essay one day. In that essay I will also explore the idea that this link with ethics introduces reasons to strive to develop an effective and authentic model of the world, rather than blindly accepting received wisdom from others.

Form and content

One concern that’s been raised with the approach that I’m taking here is that it looks as if I am confusing the fact that we all have different beliefs and values, different ‘content’, and yet we might all still reason using the same ‘form’. Don’t we all use deductions and inferences and wouldn’t all communities recognise a contradiction? So this more standard view of reason would argue that the form of reason is objective and constant but only the content varies between communities and individuals.

I think this neat partition into form and content is itself problematic. Firstly, if the form of reason was indeed objective, constant and formalisable then I believe that by now the goal of finding a good formalisation of day to day reasoning would be more successful than it has been. Note that the formal symbolic logics in Principia Mathematica or predicate logics are best understood as being models of mathematical reasoning, rather than models of reasoning in general.

Secondly I think that the form of a good argument is itself something that we can reason about and therefore the form is itself a type of content. Of course in one particular context a group of people might agree what the form of an allowed argument is and thereby define what is the accepted ‘dialect of reason’. For example, we effectively do this in law courts and in many academic disciplines. As hinted before I think the term ‘rational’ should only be used for reasoning that is adhering to a specified form in this way. However, there is no such universal agreement that covers all reason between all people.

Indeed, this kind of local agreement over the form of reasoning in a given context highlights the way that individual people might participate in multiple ‘dialects of reason’. For example, someone might be a lawyer, a member of a church and a member of a traditional tribe and in each of these parts of their lives they may regularly use different ‘dialects of reason’ that are not always compatible with each other. It’s not that their basic beliefs and values would be changing in each of these contexts, but the form of their arguments would be.

Thirdly, it’s fairly clear that different cultures around the world take different elements of form to be more or less important. For example it’s easy to imagine a culture where the words spoken by their leader are considered to be self-evidently true even if they contain contradictions over time. The mere fact that some members of that culture would be able to understand the concept of a contradiction does not mean that they share the same form of reason as us as they weigh the importance of contradictions differently. In particular, they might not consider such contradictions as undermining the veracity of the leader’s proclamations.

Finally, the approach to reason that I am suggesting here has a perfectly clear explanation of why it is likely to be the case that most, if not all cultures will share certain structural similarities between their dialects of reason. This happens because we are all trying to make effective models of the same world. In particular, deductions, inferences and contradictions are all likely features that will emerge as useful parts of any competent model of the world as they reflect certain types of consistencies in the world.

Then, not only is the form a type of content but also the source of the content matters. It’s not just important that we believe something to be true, it’s also important why we think it’s true. The assumptions (‘content’) that underpin one dialog are usually themselves the result of some previous dialog. So, while a bounded formal system does not question the origins of the assumptions being used, in day to day reasoning we often do (or should do!) which reveals the unbounded nature of general reasoning.

Why does this matter?

In some ways this entire essay could be seen as just pressing a seemingly pedantic point: reason is not an attempt to be objectively logical, rather reason is our attempt to express our model of the world. We may happen to sometimes use logic as part of our reasoning, but this is only because it can be a useful tool for modelling the world. So why does this distinction matter?

Firstly this distinction has helped me to better understand my own use of reason and come to terms with how shallow the logic behind some of my own reasoning has turned out to be when I have explored certain subjects in more depth. This in turn has revealed how hard it would be to develop a truly consistent, singular personal model of the world even while that still remains something I consider a worthy goal to work towards.

Secondly, this subjective view of reason makes it easier for me to understand why so many other sensible, intelligence, well meaning people in the world appear to me to have very poor reasoning in relation to certain subjects. It’s not just that their assumptions and values are different from mine, but the very structure of their arguments are unconvincing to me as they are engaged in a different dialect of reason.

Identifying the ‘dialect of reason’ concept allows a distinction to be made between arguments happening within a given dialect as opposed to between two dialects. This is similar to Kuhn’s notion of dialogs within and between paradigms of science. No amount of evidence or facts or logic presented within one dialect will be enough to convince members of the other dialect that their arguments are wrong.

So, the third important benefit of looking at reason in this way is that it gives some signposts as to how one might go about engaging in dialog between different dialects. Some people will see a given argument as providing good reasons, when others see it as just being good rhetoric and others still will see it simply as bad reasoning. To talk effectively with someone with a different world model you need to understand how their view of the world is shaping the way that they reason.

A forth benefit is that it helps understand why we are all able to reason effectively despite having many apparent contradictions or inconsistencies in our views of the world and the ways that we act.

Indeed, sometimes people search for reasons for something they have become convinced is true, despite knowing the contradicting evidence and logic that is currently in front of them. For example, scientists often work for years getting negative results while still being convinced that their theory is correct even though they haven’t yet got the evidence to back up their belief. Sometimes they subsequently find the evidence, sometimes they don’t.

Fifthly, I think this more behaviourist and semi-mechanistic notion of what reason is provides a much easier way to envisage how a gradual improvement in the ability to reason evolved long before humans started to understand and value the use of symbolic logic.

It’s easy to see that our nearest relatives, such as chimpanzees, have some behaviours that suggest they also have some level of ‘model’ of the world. Reason is just the ability that we humans have further evolved to use language to reflect on and express aspects of our models to each other and to judge the quality of other people’s models. This immensely useful skill helps us learn from and share the best bits of each other’s models of the world.

Finally, I have long been wondering how it might be possible to take the concerns of post-modernism seriously without undermining the value of reason. I believe the approach to reason presented here might have achieved this. It takes very seriously the cultural influences on each of our dialects of reason and yet also grounds this reason through our attempt to effectively model our shared world. So while it is a subjective notion of reason it should not fall foul of unbounded relativism.

Furthermore, by seeing reason as a particular kind of dialog that helps us better understand the world we share it helps highlight the value of reason. We certainly need to understand our world better. We certainly need reason.