Definitions don't need to be observer independent. For example, the Cambridge Dictionary defines beauty as "the quality of being pleasing, especially to look at, or someone or something that gives great pleasure, especially when you look at it"
I agree that one only knows that coffee has a strong flavour after drinking it, in that the drinker reacts to the taste of the coffee. But even so, is it still not the case that the coffee has a strong flavour, not that the coffee causes a strong flavour? The drinker of the coffee discovers a property of the coffee. — RussellA
I would say that it depends on perspective, and more generally how the given term is used.
It is certainly the case that one often uses language tautologically, as for example in the case of private perceptual judgements. For example, ordinarily I might judge my socks to be 'white'. In this situation I am using 'whiteness' to mean
my experience of my socks - I am not estimating
their colour as being the effect of a hidden-variable that is a theoretical term of public discourse, e.g. 'optical whiteness' as referred to by Physics - rather i am defining
what "whiteness" is
in my judgemental context.
The interesting thing about continuations, is that they seem to accommodate such private analytic judgements. Take the continuation
Whiteness :: For all r, (whiteStimulus -> r) -> r
The intended meaning is that the public meaning of 'whiteness' is the hypothetical set of outcomes that might occur in response to anything acting upon a particular class of stimuli called "whiteStimuli'", in any conceivable fashion.
Then take the function (whiteStimulus -> r) to mean Bob's private interpretation of a 'whiteStimulus'. From Bob's perspective, it is tautologically the case that a 'whiteStimulus' is indeed a 'whiteStimulus'
By inserting the identity function id :: white-stimulus -> white-stimulus into the previous continuation, we get
Whiteness id :: white-stimulus
We can think of the term (Whiteness id) as representing Bob's private understanding or use of the public definition of Whiteness, which as shown, is indeed is of type 'white-stimulus'.
So the public definition of whiteness as a continuation isn't in contradiction with the subjective 'private language' use-cases of whiteness by each speaker of the linguistic community, but accommodates them in the same way that it accommodates the objective physical definition of 'whiteness' in terms of the physical responses of optical estimators,.
However, continuations seem to present the problem of infinite regress; for what exactly is the definition of the type called 'white stimulus' here? presumably in some use-cases, such as in physics it is taken to be another hidden variable that is another continuation.
White-Stimulus :: For all r , ( someType -> r) -> r
Whilst in other use-cases, such as Bob's perceptual judgements, it refers to a 'given' of experience that is decided by tautological judgement.
Continuations obviously aren't the whole story, nor even necessarily part of the story for there are problems, but they seem useful in conveying the open-ended, counterfactual and inferential semantics of terms as well as accommodating the differing perspectival semantics of individual speakers.