Semantic Possibility

Part I of IV – Language, Logic, Mind, and Modality

Dr Wolfgang Schwarz, Edinburgh University

 

Intro

Semantics relates meanings to expressions through functions from which you get truth-values.

There is disagreement in semantics about what is being done in semantics.

 

There are philosophies about the meanings of different words… but what do they say about the world? What claim is being made?

—> formal semantics is an empirical discipline, like biology and unlike maths.

 

In history of logic, there was a shift in 1920s, especially in formal logic. In modal logic, logicians developed S4, a collection of statements supposed to be valid according to some values and axioms. With the shift, people found the way to understand logic in terms of models. Kripke made advancements on systems.

 

Some argue that semantics aims at systemising and illuminating facts about logical entailment (of the sort if A then B). So the formal study of meaning is really the study of entailment.

 

Strategy:

• define a class of mathematical structures; for each structure:
  • Assign to each atomic expression a part of the structure;
  • Recursively define for each sentence whether it is true in structure
  • P entails Q iff Q is true in all structures in which P is true

 

—>  But: “this is a narrow view of what semantics is because entailment is narrow.”

—> Response: sure, we don't just systemise about entailment, which isn’t a brute fact, we systemise people's judgments about what people entail.

—> This is still narrow.

—> Other responses: maybe semantics tries to uncover how our brain computes meanings. Or… judgements about those meanings (for example: the brain computes that 'every cat meows' means this, whatever it in fact means)

 

 

Semantics as a special science

The relevance of theoretical physics to linguistics: everything is just swarms of particles (materialist approach). So every process you see can be explained in terms of protons, etc, interacting. Everything has a physical explanation. So how come we don't give physical explanations to everything? How come is there biology, linguistics, politics, sociology, etc., since everything is physical?

 

For example, with an ice cube, whether it melts or not depends on how it started off. But almost all ice cubes will start off in a configuration where they will melt.

At the moment, we live in a world where protons, neutrons, and electrons etc., are arranged in very nice ways where there are things are very predictable (for example a pen falling to

the ground because of gravity). There are patterns in how everything moves around, patterns like if you hit your hand on a table, it won't go through.

—> there is fundamental physics, and then every other science, which I shall label under the category of "special sciences".

 

Special sciences study high-level robust patterns forming out of combinations of particles (thermodynamics, society, behaviour, language…) in an ultimately microphysical world, which, in theory we could explain in physics, although the calculations would be way too complicated to use for everything.

 

Many interesting patterns in human communities involve the production and perception of sounds and scribbles, and how they relate to other things people do.

 

For example, take the sentence "The building is on fire". The moment when people say that is not just random. These sounds usually occur when there is a fire. Or, "Can you pass the salt, please?", this is usually followed by an action to give someone some salt. And this is way too complicated to explain in terms of particles. So, we can explain it in terms of semantics: seeing what those sounds mean.

 

Semantics thus describe these patterns. For example, meercats have different alarm calls for different predators, since they require different protective reactions. It is interesting because making an alarm call increases your odds of being detected (curious biologically). And you can predict under what circumstances these sounds are to be heard, and what actions are  likely to follow from them.

We can capture an important pattern in meerkat behaviour by assigning truth-conditions to (types of) alarm calls.

This, in formal semantics, just describes an interesting pattern in meerkat actions. It is a summary of how particles work (n.b. ‘iff’ = if and only if).

 

 

• people produce 'it is raining' only if it is raining; and
• people respond to 'it is raining' in a manner that's appropriate if it is raining

People's judgments about meaning, truth or entailment are at best indirectly relevant.

 

When a new science is developed, new concepts need to be created, and they are given a precise meaning to attempt to describe patterns of particles.

 

—> semantics, study of meaning, is a story of what happens in the world. As a material metaphysicist, there are no immaterial facts.

 

In humans, there doesn't seem to be that automatic response like the model above describes. We thus need to add psychology.

 

 

Attitudes

Many patterns in human (and meerkat) behaviour can be usefully systematised in terms of beliefs and desires.

 

By talking in terms of beliefs and desires, you can make the prediction that because I live in Edinburgh, I will take the train tomorrow to return, a prediction which would be too complicated to make by analysing particles.

 

(Bob believes it's going to rain - his Credence in raining is 0.9)

 

What does this say about the world?

—> Bob is in a state that disposes him to act in a manner that's appropriate (given his desires) if it is very likely raining.

 

Plausibly, the production and reception of sentences is connected to behaviour via attitudes.

 

Now on the sentence "it is raining", true iff it is raining, we can say: there is a convention (well known regularity sustained in some ways) to produce 'it is raining' only if one believes that it is raining. It is arbitrary, but since everyone uses that convention, it's easier to keep using the convention.

 

So, if one hears "it is raining", then if they believe the other is trustworthy, they will react in a way that is coherent to their desires, considering the rain.

 

 

A Ramification

Philosophers of language have divided at least from the 1970s to 2000s into descriptivists and direct referentialists.

 

They have asked themselves what things such as 'the morning star' mean. For the direct referentialist, it is Venus corresponding to the morning star. Whereas for the descriptivist, it describes something captured by the way in which we describe Venus, e.g. different from another of its descriptions, the evening star.

 

But what does this say about the world, about the particles, since there only particles exist?

The question really is this: is there a convention to produce 'The morning star is a planet' only if one believes that Venus is a planet?

—> descriptivism here comes out better. Because if someone utters: "the morning star is not the evening star", we wouldn't assume that they think "Venus is not identical to Venus" (hence why direct referentialism works less well).

 

Questions and Responses

 

Q1) what is the relationship between science and philosophy?

Response: someone said philosophy of science is as useful as ornithology is to a bird, but I don't think so. It might seem annoying that philosophy explains what other people do. Philosophy doesn't have a subject matter in fact. It doesn't give conclusive proofs, and doesn't have a methodology to draw truths from inference about evidence. It's a tool-set more like systemising, drawing connections or weighing arguments for or against things.

 

Q2) what is your ontology on language?

Response: the usual philosophical tools: making distinctions others don't, finding gaps in arguments, etc. It's useful because we're trying to find patterns in the noise of the world. Philosophy is there to see patterns that other sciences don't see, for instance the relationship between sciences. And this have often led to the birth of news sciences. Ontology in the strict sense is the theory of what there is, what exists, and my ontology is that of the mere physical world. Language and meaning also entirely exists, but not as anything extra to the particles. So once you've settled all the particles, you don't need to add anything to get our world, you have our world.

 

Q3) what do you make of patterns?

Response: often regularities.

 

Q4) is a pattern something which is psychologically satisfying?

Response: some of the regularities seem to us to be worth studying. Are they objectively privileged or do our own interests (what we find psychologically satisfying) determine what we study? I somewhat agree that it is what we find psychologically satisfying, but there is a limit. For instance, physics does seem to be saying more than just what attracts me about regularities in the world. Laws of physics seem to depend on more than just on our interests, though the studying being based on our interests might be more true in other sciences.

 

Q5) About the idea of pushing one’s hand through a table: Newtonian mechanics comes out of that at a higher level than particle physics, and I wondered if semantics too?

Response: Yes, much higher level. If you look in detail at what's happening when I put my hand through the table, there is in fact no certainty that my hand won't go through. But patterns in more than 2-3 particles are messy and noisy, hence why we study in special sciences, such as semantics.

 

Q6) I am interested in what you said about what semantics tells us about the world. I see a lot in physics, and other ‘harder’ sciences, about what it tells us in the world, but harder scientific theories also conjecture why things work the way they do. Can your approach to semantics offer the ‘why’s, rather than just the ‘what’s?

Response: Yes, potentially… and it might be what we measure with the robustness of patterns. When you explain an event, you at least in part see how it depends on various other things (for example: would it have happened if we removed this or that fact or state?). But there are different kinds of explanations, for example mechanical explanations are more about how it works, and not why. In semantics, thanks to the regularities, we can have an explanation, but it's unclear what the 'why' would be… unless perhaps why one says something, but my presentation earlier sees it more in a descriptive sense. The brain could be an explanation, though I still want to reject the idea that semantics aims at explaining what the brain does.