Sets are typically defined in terms of their characteristic properties. The set of human beings is characterised by whatever properties human beings share. In linguistics, a phoneme is defined as a set of allophones which are more phonetically similar to each other than to other allophones (and which differ in terms of their positional distributions in words).
The same set can be defined by different properties: The set of beings which have kidneys is (apparently) the same as the set of beings which have livers. The set of lexical items which have forms is the same as the set of lexical items which have meanings.
The characteristic semantic properties of words are known as semantic components - sometimes as semantic features or semantic markers of words.
This is a rather simple view of the matter - componential theories of semantics are often reduced to attaching labels like [+ animate] to words with similar meaning, without explaining exactly what these labels themselves mean.
Considering that meaning is a relation between words and the world, in very general terms, this is a rather circular approach, which effectively paraphrases words in a subset of English which is used as an artificial language.
However, componential analysis is a useful heuristic technique which can support more sophisticated semantic analysis in various ways. The classic areas for componential analysis are:
Tasks: