To this day, I have not been able to envisage a sound answer to this question and, to my mind, this is (one of) THE pending question(s) in linguistics. Any insights will be more than welcome.
The meaning of the term 'context' is broad, and varies among academic work. Context may include the following features:
-Linguistic context: the stretches of discourse that precede and follow the stretch of discourse to be analyzed;
-Situational context: the place and time in which the stretch of discourse occurs; the persons who take part in the speech event; the surrounding objects and events taking place simultaneously (for example, if two people are speaking while fixing a lamp, this action is part of the situational context). This context includes the knowledge of the world that each participant in the speech event has and assumes the others to have (in this respect, a conversation among Spanish students is different from a conversation among students of different countries who have recently arrived in Spain). The situational context also includes related previous entities and events: for example, context elements of an essay on Oscar Wilde include his life, works, the historical period in which he lived, and also preceding academic work on this author.
There are interesting books on contexts by Teun van Dijk. Moreover, I suggest a paper: E. Rigotti - A. Rocci (2006), Towards a definition of communication context. Studies in Communication Sciences 6(2)
In experimental research, context is developed in many ways.
One common division in human research is between immediate context and general context. How comfortable the chair you are sitting in is immediate context. What room you're in is general context. This can be altered by influencing attention, though.
And that's really the main distinction with regard to context being operationally defined. Something that is almost definitely directly affecting attention is different from something that is unlikely to be directly affecting attention.
Then you can start working with different types of attention. You may not consciously notice whether the room you are in has a clock or not, but you may non-consciously notice it.
In general terms, every research study deals with context at various levels. Your dependent and independent variables can be reframed as types of context. You generally report on the demographic characteristics of the subjects, such as gender. This would be a type of context. Usually one that you hope has no effect.
Most environmental (which computer is which subject using?), demographic (does it matter that almost every experiment is done on university sophomores as a psychology class requirement?), or individual state of being (who's hungry?) elements of context are things that you hope will have no effect or that the differences will cancel each other out.
Animal research is much more restrictive in terms of controlling both the environment and the differences among test animals. This is a way of excluding unwanted contextual influences. And what is one experiment's unwanted and excluded contextual influence is another experiment's subject matter. People have conducted research on populations of fruit flies in various types of context (environmental and phenotypical) for up to 50 years; tens of thousands of generations selectively exposed to different elements of context.
Biomedical research also has differentiation of type of context that range from extremely carefully measured to sparsely reported and incomplete. If you look at the regulations governing the development of pharmaceuticals for human use and the reporting requirements imposed after a pharmaceutical is available you will find many layers of context implicit or explicit within the regulations. Which, in turn, correspond to human actions, linguistic and otherwise.
Again, though, this can be divided into things that get direct attention, things that intentionally get indirect attention (or for which vigilance is expected), things that non-intentionally get indirect attention, and things that get no attention. And regardless of where something falls within those categories, it can also be something you hope doesn't get attention.
Attention to the environment is as old as life itself. Language is about attention to an environment. Context is what gets attention or affects the capacity for attention that is not the main goal of the entity at the time.
Context in my mind means 'place'. Things commonly will only have meaning if they are seen in place. Archeology is a case in point. When an artefact is found in situ it is in context in place. It means more to the finder when they see it in that place than it would if handed to them in a bar in downtown Manhattan. A good deal of archaeological knowledge has been lost because 19th century archeologists thought the removal of the artefact to a museum was more important than the observation of it in 'context'
In my current world the places or contexts are in vitro and in vivo. The reaction of a tumour cell to the presence of a drug can be observed in vitro, that is a scientific place or context. The action of the same drug on cells in vivo is a clinical 'place' or context.
Context does matter a great deal in looking at archaeological artefacts in situ and cells in people. Artefacts in museums and cells in flasks might be 'out of context'
This strikes me as an interesting question (though I have just come across it), and I particularly liked Turner's answer as providing a useful example of the role of context in meaning, significance or understanding. Removing an archaeological artifact from the place where it is found is likely to diminish the attempt to understand its significance, since its place and surrounding objects or the surrounding situation may well provide important clues.
Looking at the matter more strictly, as a matter of language or linguistic context, appeal to context (and notice the morphology here "con" + "text" = what is with or together with the text) is regarded as important in criticism of the journalistic practices oft connected with "sound bites." The idea is that the short, pithy sound bite, though it grabs attention, is often misleading, since in order to properly understand it, one must know what else was said. Providing needed context is then a matter of more adequate reporting. Selective reporting which ignores needed context is often a sign of prejudice or deep prejudgment, excessive partisanship, etc.
There is, again, the famous "context principle" due to G. Frege, a philosopher who did basic 19th century work on meaning and interpretation. Frege's context principle goes like this: "Don't ask for the meaning of a word, except in the context of a sentence." (Compare, e.g., "He put his money in the bank, and a bank is the side of a river, therefore, he put his money in the side of a river." What goes wrong here?)
In good contemporary dictionaries, the principle is often observed, by making separate sub-listings of the various meanings of a word, together with illustrations of related, distinctive usage. This is especially useful in translating dictionaries, where the novice learner, would otherwise simply guess at which word to select from a listing of possible translations of a word in the home language.
Frege's context principle is usefully compared to related views of Wittgenstein on language and meaning. In particular, later Wittgenstein is often quoted to the effect "Don't look for the meaning, look for the usage." This combines his skepticism regarding purely referential accounts of linguistic meaning (as though the meaning of a word were an object of introspection, e.g.) with due emphasis on usage.
When dictionaries are written, this work rests on wide-ranging empirical research which involves the collection of examples of word usage--sentences or short texts. Looking into the famous Oxford English dictionary (OED), for example, one will often find various examples of the usage of a word reaching back over hundreds of years. The lexicographer then takes this empirical evidence of usage, and will formulate it into definitions which seek to summarize examples or ranges of examples of usage. Particular illustrations of usage are then possible counter-examples to proposed definitions, though given serious, deeper ambiguity, we may instead end up with different dictionary entries, or sub-entries, for the "same word" phonetically or orthographically identified. From this perspective, Wittgenstein's dictum, like Frege's, is good advice on questions of lexical interpretation.
All of this strikes me as relevant to the understanding of many inquiries and debates in scholarly contexts, concerned with the interpretation of classical or other important texts. In such cases, the appeal to context is often quite broad, and we are asked to examine the details of usage in an author's text of interest, and perhaps also the usage by the same author in other, related texts --or even contrasting usage of related authors in entirely different texts. There are many possible illustrations in related practices of the point that adequate interpretation and understanding rests on the proper consideration of relevant context.
The term "taken in context" is one of the most difficult phrases for me to understand. What context, were are the boundaries, and in what type of context should I use for this particular set of signs?
To me context is specific and should be illuminated in all published papers. In other words context can be both a bias or ideology as well as a just scientific lens in which to view data.
Context is as Foucault describes a discursive field. This discursive field has limits and is articulated in The Archaeology of Knowledge and the Discourse on Knowledge (1972). Just as Barry suggests place I would agree with Foucault as he describes placing in or linguistic place or relevance.
Take for example four people setting around a dinner table having a conversation about fashion. I could be included in the discourse surrounding fashion as long as it was in reference to appearance. Yet once it comes to him lines, types of fabric, what was hot last year I am excluded from at worst or not understanding the discursive field or context at best.
This is a major part of what is labeled as critical discourse analysis and as has been mentioned van Dijk and Fairclouth along with Foucault are the fundationalists in this linguistic field.
Simply put the context is the limits of one discursive field and the beginnings of another.
To me context is specific and should be illuminated in all published papers. In other words context can be both a bias or ideology as well as a just scientific lens in which to view data.
--end quotation.
Where context is an aid in understanding, I think you are right that the needed context should be specified as finely as possible. It is also true that a misleading context may be a "bias or ideology" as you say, as well as "a just scientific lens in which to view data." The refusal to provide needed context can also function in the service of bias and ideology. That's the most frequent criticism of mere "sound bites" in journalism --that it easily amounts to a kind of biased selection.
One mark of a bias is that it tends to be limiting. It might be better to think of this in terms of "pre-judgment." People tend to bring their pre-judgments or the collection of their prior judgments with them, as they encounter new information. It strikes me that a bias is an emotionally charged kind of pre-judgment, which tends always in a single direction and which is not responsive to facts and new information. Pre-judgments of the wise contribute to their wisdom, while the prejudices of the narrow-minded limit their further growth --moral and intellectual development.
In contrast to mere bias, in the usual every-day sense, "ideology" tends to be something quite elaborate. It is a mark of ideology, however, that it tends to be a bit too neat for wider intellectual respectability. The typical ideologist has an answer to every possible objection, however contrived, and discussion with the ideologist tends to go around in circles. Moreover, people of the same ideology tend to all agree on every essential issue; and they will often form a kind of inward-looking tribe, since they are most successful and most intent on conversation with each other --where they supply a great deal of mutual emotional support. It can be difficult to identify ideologies, but once they are identified, then they will usually prove unsuited to the role of supplying suitable context, even where conducted in the same vocabulary, and even where the same issues are in question.
A proper Contextualism supplies grounds for criticism of bias and ideology.
Much more could be said on the idea of context as a "proper scientific lens."