What is the Distinction Between Sense and Reference?
Frege is said to be the first person to set out the difference between sense and reference in a systematic form, and it is from his writing that the terms first arise. It must first be noted that since his works were written in German there is a small amount of debate on how the terms should be correctly translated. Dummett suggests that the term Bedeutung might more properly be translated as simply 'meaning', but he also points out a difficulty with such a translation - that Frege seems to be making the case that a word's reference is just one part of its overall meaning.
A reference must be a public, neutral object - one that does not exist merely in a person's mind, and that has existence without any association with a perceiver. A sense, on the other hand, is a description given usually in an attempt to point to a particular reference. When talking about a particular sense of a word, therefore, there need not be an associated reference, since it is possible to describe something well enough that a listener might have a pretty good idea of what something is, and yet there still not be an existent public neutral object to which the sense refers.
A good example of this, and one which seems to be frequently used, is that of the unicorn, whose sense is easily expressible - a horse with a single horn on its nose, and perhaps magical capabilities - yet there does not appear to be a reference for such a creature, for, whilst there is a description, there is no public, neutral unicorn in existence. Likewise with the case of 'magical capabilities', we can imagine what such things would be like - with strange and unexpected occurrences seeming to occur in correspondence with the will of the particular creature wielding the capabilities - without such things existing. Some might try to claim that it is not in fact possible to talk using words which do not in fact have a reference, and they might explain that in talking about a unicorn we are in fact referring to the literary references to unicorns, and that through such references unicorns do exert some sort of existence. Dummett disagrees with this claim, and I think he is right to do so. We are not simply talking about a literary convention, allowing people to know what unicorns 'really are' (or otherwise what we mean when referring to the conventions and descriptions themselves?), but instead we are talking about them as if they were real, in just the way we are able to talk about cats since they are real. Senses, therefore, can be imagined or invented, though references, being independent of people, clearly cannot.
Russell uses the example of Hamlet, saying that we do not assert in watching the play that a Dane actually called Hamlet really existed, and such an assertion is unnecessary for one's enjoyment of the play. Even more importantly, however, it would be ridiculous if someone appealed to a Dane that in fact turned out to have lived and be called Hamlet saying things like 'Hamlet was never suicidal so Shakespeare got it wrong'. Clearly there can be no effect of a real life Hamlet on the character in Shakespeare's play, and even if one exists he can't have a bearing on the play - this would seem to imply that the reference of the character in the play cannot be someone who actually lived, but merely an invented person in Shakespeare's mind who never in fact existed. Whilst the sense may be the same, the claim cannot be made that simply because the character is a historic Dane who was Prince that that is the character's reference, and clearly if we did make such a claim some very mistaken conclusions could be reached.
If you are to take the view that idealists, such as Berkeley, hold you would probably claim that, since things exist only as concepts in people's minds, and have no existence apart from their perceptions, that there are in fact no references. It is not necessary to deny the existence of references (though I imagine most idealists would prefer to do this) as it is possible to suggest that whilst there is such a thing as a reference in fact no words actually have one. This would seem, however, a rather pointless claim to make, and for most purposes would seem to be taking the side of the idealists.
As Frege puts it 'In grasping a sense, one is not certainly assured of a reference'; basically stating that just because we are able to conceive the idea of a thing does not mean that we can necessarily understand the object that this idea points at. The example that Frege uses, and one that is endlessly brought up again, is that of the 'evening star' and the 'morning star'. The reference of these two names is, in fact, the same, since they both refer to the planet Venus, and yet there is nothing intrinsic in them that requires them to be the same, since the terms have different senses. Since we are able to hold different thoughts about different senses pointing at the same reference it would be perfectly possible to believe that the 'morning star' were the planet Venus, whilst the 'evening star' were not. There might, however, be problems created by a failure to identify these two terms with one another, and with their mutual reference, since, for example, the statement 'On an equinox day on the equator the 'evening star' will be seem approximately twelve hours after the 'morning star'' would be incorrect, since evidently the 'morning star' will be seen at exactly the same time as the 'evening star', what with them being the same body. Such a claim, however, would mistake the uses of the sense and reference, since, it might well be claimed, the 'morning star' is, by definition, the star that you see last, at the beginning of a day, whilst the 'evening star' is the one you see first, at the beginning of a night, and thus you can't claim that you could ever see them at the same time, but only see their reference.
Frege debates, and Dummett picks up on this, when it would be possible to change the sense in a statement, without altering the truth value of the sentence. It should always be possible to change a sense without altering the truth-value of a statement; in the cases where this does not seem to be so it is likely because in that particular instance the reference is in fact the term or the sense. In a logical re-writing of plain English sentences such a case is usually, but not always, indicated by inverted commas around the term in question. Obviously with a sentence like ''Amanda' means 'having to be loved', in Latin' it is not possible to substitute the name Amanda for another sense pointing to the same Amanda that you know as a reference. The sentence ''The girl I know, with blonde hair to her shoulders, who wants to go to Spain to work to escape the cold' means 'having to be loved', in Latin' clearly doesn't mean the same (and in this case doesn't seem to mean very much at all). The reason for this is that, in this particular case, the sense - 'Amanda' - is the reference itself. In another circumstance, where it was the girl herself that was the reference changing the particular sense used should not alter the meaning of the sentence. 'Amanda is not intending to go to university' means the same as 'The girl I know, with blonde hair to her shoulders, who wants to go to Spain to work to escape the cold is not intending to go to university', since the reference to whom they both point is the same.
Frege seems to deny this, suggesting that the statement 'Columbus inferred from the roundness of the Earth that he could reach India by travelling towards the west' is very different from 'Columbus inferred from the roundness of the planet which is accompanied by a moon whose diameter is greater than the fourth part of its own that he could reach India by travelling towards the west'. It is not clear what the objection to this change is. If the important element is that it does not identify the planet with this particular moon with the planet on which Columbus was sailing then the objection is fair - since it is the fact that he is on the particular round planet that is important, and the planet with a moon with a diameter the fourth its size could be a completely different one. The problem in that case is that by changing the sense the reference that is pointed to is changed, or at least made ambiguous. Frege might mean, however, that because Columbus did not necessarily know about the size of the Earth's moon he did not have any beliefs about 'The planet with a moon whose diameter is a fourth its own'. This is clearly not the case as, unless 'The Earth' were rendered (quite unnaturally) in inverted commas, we must presume that it is not Columbus' exact thought that is being quoted. In that case whatever his conception of the Earth's place in the universe, or of its satellites, since a contemporary reader of Frege knows which planet is meant by the description it is taken that it is only Columbus' thought that is being preserved, and his references but not his direct words or sense.
If senses can happily be exchanged for one another and would never change the truth-values of a sentence, then having more than one sense for any reference would seem totally superfluous. This might easily be the case, but Dummett offers an alternate theory, and one that is quite attractive. Dummett suggests that there is a third part to the meaning of any word, and that, in fact, the meaning of a word is made up of sense, reference and tone. Tone is in fact much more easily understood than the other two properties of meaning, and simply refers to their resonance, and the evocative aspects of the particular word. It is easy to understand using Dummett's example that in poetry different uses can be found for the words 'dead' and 'deceased'. Clearly it is not just in the case of poetry that particular words are chosen for their effect, however, and the choice between legal vocabulary and colloquial speech, for example, will play a large effect, as will a large number of other factors.
A problem that arises out of tone, however, is that Dummett feels that many other philosophers would like to put it down as subjective, with the evocations meaning different things to different people, and he cannot understand how this could practically work, since surely meaning must by definition be objective, and language is used to convey opinion and meaning. The simple answer is that of course tone could not wildly change the usage of a word (and if a particular speaker supposed it did he would end up using words incorrectly, and would not convey what he meant to), but surely it is possible that slight fluctuations may be made using the tone of a word, such as whether the statement is presented in a pejorative manner or not. Without changing the truth-value, tone could be used to affect how it is put across. Obviously this would not entitle people to use words entirely subjectively, but this would allow people to use them slightly differently. It can often be clear in a context why a person has used a particular word, and in this sense the tone would have to be objective (or it would not be possible to establish why they chose a particular word), but the choice itself would reflect someone's personal opinions.
Overall, therefore, Frege's 'discovery' of sense and reference seems a very important and useful one. Dummett tries to claim that there should be no distinction, yet I think does not make a very good case - and it always seems strange to argue with someone who has defined terms, trying to claim they haven't defined any separate things. Dummett does make a contribution with his ideas about tone, which, whilst not necessary, do add to the completeness of the theory. It seems very hard to define sense and reference without an appeal to examples, but if I were forced to do so I think I would simply say that the sense is what someone is trying to express with a word or phrase and that reference is the public, neutral object to which this expression alludes.