[WS:] I think the question of "value" is an equivalent of the theological problem of counting angels on the head of a pin. Like angels, "value" is a mental phenomenon that not only is a product of social interaction but it also has multiple, impossible to separate in practice meanings that range from measurable (i.e. price) to immeasurable (e.g. emotional values.) For example, the ownership of an object (social relation) affects the perception of its monetary value, which is known as the endowment effect http://en.wikipedia.org/wiki/Endowment_effect.
Needless to say that the perceived social desirability (or lack thereof) affect not only the price but also other types of perceived values of an object. A good example is the value of US made blue jeans in the USSR (and other Eastern European countries.) It was perceived as something very desirable to have because it bestowed certain social status on the owner. This affected not only the price (about one month's wages) but also the "use value" (i.e. people wearing the garment regardless of its suitability for the occasion) and the emotional values attached to it. These values were totally unrelated to the Marxian (Ricardian) concept of value as socially necessary time to produce the garment. Likewise, the cost of production of, say, an Ipod is a fraction of the price it fetches in the US, or the value attached to it by the fans of this line of product.
>From my pov, the question what determines value is fundamentally a
metaphysical one (in the sense defined by Kant in "Prolegomena to Any
Future Metaphysics"
http://en.wikipedia.org/wiki/Prolegomena_to_Any_Future_Metaphysics) -
i.e. a question that we ask because asking such questions is a natural
disposition of human reason, but it is not possible to answer it by
rational means (i.e. as we answer questions in science.) We can
endless debate what it is and what it is not, but there are no
effective means of determining it. The only exception is the question
of price. price is an empirically observable phenomenon, and its
value can be estimated through empirical generalizations that are by
definition provisional. That is, its values can be statistically
estimated by the means of correlation with other measurable values,
but as we all know, correlation is not a proof of causality.
Wojtek