Issue #27 December 2019

Why Thoughts Are Not In The Head: Frege on Sense

“So the result seems to be: thoughts are neither things of the outer world nor ideas. — A third realm must be recognized.”

— Frege, 1918, p.302

Where are thoughts? This question is not that one on the longer road to the problem of consciousness: ‘Thoughts are seemingly ‘in’ our heads in a way different from how our brains are ‘in’ our heads’. But that’s not the question I want to pursue here.

Let’s assume that it is unproblematic to speak of a mental ‘realm’, in addition to the ‘realm’ of the world, that here are two cartographies, as we do when we commonly speak, day to day, about things. Now, with these assumptions in place, let’s repeat the question “Where are thoughts?”. The answer might seem obvious and trivial — they are in the mental realm, wherever that is. Thoughts are ‘in the head’ as opposed to ‘in the world’. In the world are just objects and forces, and their context, and then there is us, thinking about it, having thoughts about this or that part of the world that lies around us. However, problems begin to creep into this seemingly intuitive account.

Let’s begin unpacking what already our common sense ontology has given us. There are mental and objective realities. The objective reality is publicly shared, whereas the mental is private and individual. There are thoughts. These thoughts can be true or false depending on what is happening in the public, objective, world. There’s also language. Language permits of sentences that can also be true or false in a similar way to thoughts. We might want to say, therefore, as we often do, that language expresses thoughts. Thoughts are private to the individual, who makes use of language to express them, and if the thought is true in reference to what is happening in the public, objective, world, then the sentence that expresses it well will likewise be true. This seems to be the common, ‘folk’ story of what is happening with all of these things.

However, problems begin to emerge when we ask how we can say if two thoughts (and, by extension, two sentences) are in fact the same thought or sentence. In our just sketched picture this could be two people possessing the ‘same’ thought at the same time, or at different times, or one person possessing the same thought twice at different times. With the resources we have above we can say that the thoughts are the same just insofar as they are about the same ‘chunk’ of the public, objective, world, and make the same claim regarding it. That is to say that they are linked such that the very same circumstances that would reveal or make one thought true would make the other one true. If this wasn’t the case, then, perhaps despite appearances, we are dealing with two thoughts.

So, for example, if I think “Michael’s cat bites”, and you think, at Michael’s house, “that cat looks like it’d bite someone if given half the chance” we can be said to be thinking the same thought, because both thoughts turn on the identical circumstance whether or not they are true, namely, that fluffy thing over there with the teeth and the frequency with which it uses said teeth against people. We see this even clearer if we both had expressed our thoughts in language, with the first of us then being able to say to the second “That’s exactly what I just said”.

And, conversely, if Michael then says he better put the cat outside for our safety, and that he’d be “back in a minute”, so we both think that “Michael will be back in a minute”, it might appear that we now have the same thought. However, if, after 60 seconds precisely, I say “Michael was wrong. He said he would be a minute, and it’s been a minute and he’s not back” you might realize that we did not have the same thought at all: if Michael comes back in two minutes it would still be true that he was back ‘in a minute’ if we understand that in the common way. My understanding Michael’s sentence overly literally means I have a different thought from you who understood it colloquially, and the difference resides in the different circumstances required to make one or the other true (if he was back in precisely 60 seconds, both of our thoughts would be true, but for different reasons, i.e. slightly different circumstances).

However, there is an interesting puzzle that emerges here. If what makes our thoughts the same is their turning on the identical set of circumstances that determine whether or not they are true, we can then make as many substitutions in how the thought is expressed as we wish and apparently always have the same thought, just so long as we make sure the new substitutions turn on the same circumstances. If this was the case, if you ever thought something true about Michael’s cat, you also have thought everything that is true of Michael’s cat, including the barcode on the cat food it ate yesterday. Why? Because if I wrote out every single true thing about Michael’s cat, and then clustered that all together in a noun phrase “The being that …”, then I could substitute that linguistic monstrosity into the subject of the thought “Michael’s cat bites”, and have expressed the ‘same’ thought. I.e. ‘the thing that belongs to Michael and ate the cat food yesterday which had the barcode xxx-xxxxxxx bites’.

Perhaps a simpler example will make this problem clearer. Frege gives the example of the previous names of Venus: the morning star and the evening star (1892a). Hopefully the reason for Venus having these two pseudonyms is clear. If I were to tell you, just having found out about Venus as a planet and how it appears from here on Earth, that the morning star is the evening star, that is, they are one and the same, to which you replied “the morning star is the morning star, the evening star is the evening star” in what sense is the same thought being expressed once by me and twice by you? Seemingly none at all. But if there is in fact only one star, then all of these sentences turn on the exact same set of circumstances whether they are true or not. It looks like we are dealing here with three distinct thoughts:

A is A

B is B

A is B

The problem is that if it is true that the morning star is the evening star, then whether I use one term or the other shouldn’t change the ‘thought’ because the names have no bearing on the circumstances that the three thoughts turn on regarding whether they are true or not.

Frege (1892a) offers another example. Imagine three lines, a, b, and c, that run from each of a triangle’s vertices to the midpoints of the opposites edges of the triangle. Of this situation, we can say that the point of intersection between lines a and b is the same point of intersection between b and c.

Now imagine you’ve lost your keys in the grass at a park. A passer by asks you what you are looking for and you tell him. He says he passed some keys lying in the grass back a ways, and you ask him to indicate where they are. After a bit of fruitless pointing, he indicates that starting at the picnickers over there, if you walk in a line towards the midpoint between the tree and the barbecue, you’ll find them right when the tree is on your left and the midpoint of the picnickers and the barbecue is on your right. Odd directions, to be sure, but admirable for their Cartesian precision. However, when you get over to the area you have (understandably) forgotten the details of where to start and which thing should be on the left or right. Your friend, though, points out it does not matter, just pick one of the three, head towards the opposite mid-point, and stop when you’re standing between something else and another mid-point. Long story short, you find your keys, right at the intersection of ABC, you also learn the geometry lesson from your friend (that all of these lines intersect at the same point). Now, assuming the direction giver didn’t know that all of the lines intersected at the same point, and was literally giving their best way of indicating the point where the keys were, in what sense could it be said that they were saying the same thing as your friend with their more abstract ‘pick any point and…’ direction? If the only resources we have is the physical world, and thoughts, it would seem the thoughts (or sentences) are equivalent. But it seems strongly here that there is a difference.

What differs is the manner in which the thought presents itself, the ‘mode of presentation’. Both the direction giver and your friend are saying the same thing about the same keys being in the same point in space, but they are presenting this fact in two different ways, one more general than the other. The road to them is different. This is not say that one has their idiom and the other has theirs, because separate idioms don’t necessarily imply a different thought, otherwise no one would ever have the same thought (because it seems arbitrary to demarcate — do different accents change the thought?). The difference between the direction giver’s and your friend’s presentations of the point where the keys lay would still be a difference in any idiom, or even language.

Frege (ibid) refers to this ‘mode of presentation’ as sense (Sinn), and argues that it is necessary to account for the difference between statements (and thoughts) like A = A, and A = B. If A and B are in fact equal, then they are just one thing, and whether we use one name or another should (in our simple, common sense ontology) make no difference. But there is, seemingly, an important difference, thus there must be something beyond words and objective things to account for this difference.

At the level of individual names (or phrases that would act as definite names, what Russell called ‘definite descriptions’) there is a sign, that has a referent (often an object), and the sign refers to the referent via a sense, a mode of presentation. The relation here is from many to one: many signs could pick out a single sense (as when a name of a city is translated), and many senses could refer to a single referent (such as ‘17’ and ‘9+8’). At the level of the sentence, the ‘sense’ of a sentence is a ‘thought’, and the reference of this thought is a truth value. Why a truth value? The reasons for this get a little complex, and have to do with the idea of substituting equivalent terms into a sentence. If terms are truly equivalent, then this should not alter what is being denoted by the sentence. But it is possible to perform a series of substitutions that seem to draw the sentence away from one state of affairs onto another, as long as each step renders a true sentence (see the slingshot argument). This means that every true thought has the same reference: the ‘true’.

However, the key point for our purposes is that these ‘thoughts’, the senses of propositions, are necessarily non-subjective. Frege contrasts them to ‘ideas’, which he seems to mean in the empiricist’s sense, as a cluster of past or present sensory data. These, owing to the subjective and private position from which they are engendered, can never be shared. Thus, there are as many ideas as people having them. However, it is important for our consideration of the question of counting thoughts that it was possible for two people to have the same thought, that two people could think or say the same thing. A moments consideration also shows us that if that were not the case, then language could not achieve its inter-subjective function (assuming, of course, with our simple ontology, that sentences express thoughts). A group of people will generally have different ideas that are called up by the name “The Statue of Liberty” (different images and associations); if there is any coincidence between these ideas then this itself would be mere coincidence. However, this myriad of different ideas has no bearing on this group of people being able to grasp that the sentence “The Statue of Liberty is on Liberty Island” is true while the sentence “The Statue of Liberty is on Staten Island” is false. And should there be disagreement about this, then we wouldn’t say that there is a just an insoluble difference of ideas, but rather that some portion of this group were simply wrong, given the senses and objects involved.

This leads Frege to assert that we don’t ‘have’ thoughts as much as ‘apprehend’ them. Sense, though accessible to language, precedes it, and provides it its condition for providing true and false sentences. The Pythagorean theorem is a thought (a sense) that has as its referent ‘the true’ and it sits there whether or not we have ever talked about it, or apprehended its truth. This makes thought sit in a ‘third realm’ between the public, objective, shared world, and the private, subjective world. Not a concrete physical occurrence, nor a pure mental idea, but something that allows the one to communicate with the other, the ‘in virtue of which’ we can encounter each other in language and discuss the world.

Frege is mostly interested in language, its meaning and its relationship to truth, but do we apprehend senses (thoughts) in other ways beyond language, or is language our only access to it? I’d like to close with an argument that Frege is compatible with a phenomenological reading that finds in our direct sensory experience of the world the kind of thing that ‘meaning’ could be attributed to, and thus sense.

In his ‘morning star’ ‘evening star’ example, Frege presents these two names as having different senses, not merely different signs (words). The relationship between a referent and its senses is one to many, as is the relationship between a sense and the signs that denote in virtue of it. We want to say that when it was discovered that the morning star is the evening star, both of them Venus, that a discovery was made, not just about the meanings of words, but what these words denote. Discovering the morning star is the evening star is a different discovery from discovering that Venus is ‘金星’ in Chinese. What was believed to be two things is discovered to always have been one, whereas the latter discovery is purely a discovery about signs and signs only. As we’ve seen we need more than just what they denote to account for this (because what they denote was always the same thing), and thus we speak of the two names having different senses, different modes of presentation of the same object. Now, it is possible for two names to point to the same sense, but not for one name to point to two senses (unless of course it hinges on an ambiguity, thus, has two meanings and is only coincidentally the ‘same’ sign, kind of proving the point). This means that multiple signs are not incompatible with a single sense, but two senses are incompatible with each other. When I discover ‘the morning star is the evening star’ I am not expanding the senses of the two names into a complex, I am apprehending or grasping a new sense, that of ‘Venus’. However, the senses of ‘morning star’ and ‘evening star’ are still there, just not grasped anymore.

Now, what is so incompatible about the senses of these two names? Precisely in that one is Venus in its mode of presentation as a bright star seen in the early morning, and one is Venus in its mode of presentation as a bright star seen clearly in the early evening. Given these different modes of presentation we can say many further things about these two different ways that Venus has of presenting itself (for example, the azimuth and meridian coordinates of their presentations on a given evening/morning) and all of these would also be true of (could be said of) Venus. The sense of Venus would be like, in the triangle example, the point of intersection of ABC, whereas the senses of the morning and evening stars would be the point of intersection AB, and BC, respectively. In the triangle example we say the smaller ‘senses’, picking out the point on the plane, differ from each other in an incommensurable way because they are composed of different lines (even though the point they describe is the same point). But what is the difference of composition between the morning star and the evening star (even though they are picking out the same star in the sky)? Precisely the phenomenological presentation of a star seen early in the morning and a star seen early in the evening. A spatio-temporally situated consciousness is required to discriminate these two modes of presentation, they are meaningless without them (also considering that where ever the morning star presents itself, the evening star is presented somewhere else in the world).

This doesn’t mean that senses have been brought back into consciousness, that they collapse back into ideas, because the two different phenomenological presentations within spatio-temporally situated consciousness still permit of an endless array of private ideas that don’t touch upon the ‘senses’ involved. Every time I think of the phenomenological presentation of the evening star, for example, I ‘see’/remember the sky above my childhood home in the summer time — I’d be surprised if you were also imagining my childhood home. No, the result is the opposite, that the phenomenological presentation is ‘pulled out’ of the mind, and now, as a sense, sits in that realm between the private subjective and public objective ‘worlds’, allowing one to communicate with the other.

So, thoughts are not ‘in the head’ and it would also seem now that neither are the whole of ‘perceptions’.

John C. Brady is a student of philosophy and educator situated in Beijing. He gets most of his reading done in traffic jams. He is also a co-editor of this magazine, by way of full disclosure.

Works Cited

Frege, (1892a), On Sense and Reference. Web link.

Frege, (1892b), On Concept and Object. Web link.

Frege, (1918/1956), The Thought: A logical inquiry, Mind, Vol. LXV, №259. Web link.


December 2019


Self-Defense, Necessary Force, and the Ethics of Modern Warfare

Daniel Rhodes in conversation with Helen Frowe

Frantz Fanon: Anticolonial Revolutions and Revolutionary Theory

by Timofei Gerber

How to Drive a Car: A Defense of Achievement as Competence

by Christopher Carroll

Why Thoughts Are Not In The Head: Frege on Sense

by John C. Brady