All of what is said here is, of course, subject to the judgment of the professional and scholarly community. Confirmation, correction, and dispute will be necessary for any progress to occur. (p. 72)
I like this line. In one parting remark at the end of his most recent article, John Budd has shifted his burden onto other shoulders. And it’s a good thing, too, because there is a heck of a lot to think about in his brand, spanking-new “Meaning, Truth, and Information: Prolegomena to a Theory“
Setting it up
I’m very interested in the roles of information, truth, and meaning in librarianship in general and information literacy in particular. Two of the goals I’m currently picking away at are (1) establishing the lack of a consensus among librarians about what ‘information’ is and (2) advocating a particular theory of information for librarianship. So, I was a little surprised to see John Budd’s contribution to the most recent Journal of Documentation, wherein he attempts to do the exact same thing! Damn it! There goes my paper. Yet, though Budd does a great job explaining the lack of a consensus, I can’t help but notice some substantive problems with his assessment of the meaning of ‘information”. Let’s take a look.
Librarians don’t know what information is
The thesis of Budd’s paper is that information “cannot be defined unless within the context of meaning and truth” (p. 56). Awesome. I couldn’t agree more. Later, he adds that “[a] reader from within or without IS would reasonably respond to the initial challenge of examining truth’s relation to information by stating that definitions of both are required before anyone can proceed” (p. 58). So, Budd begins with defining “information”. His analysis covers all the usual suspects (Buckland, Rowley, Hjorland, Losee, etc.) and ends with the observation that in the whole of information science, there is not yet a definition that “establishe[s] parameters that enable inquiring and praxis” (p. 60). Here’s a brief rundown of theories considered:
- Buckland (1991) and the “information as thing” approach.
- Kaye (1995): Information “is a central and defining characteristic of all life forms, manifested in genetic transfer, in stimulus response mechanisms, in the communication of signals and messages and, in the case of humans, in the intelligent acquisition of understanding and wisdom” (p.37)
- Brookes (1974): information as that which modifies a knowledge structure
- Eaton & Bawden (1991): “information is a dynamic force for change in the systems within which it operates” (p.59)
- Rowley (1998): Information is a relational property.
- Losee (1997): “information may be understood as the value attached or instantiated to a characteristic or variable returned by a function or produced by a process” (p. 267)
- Bates (2006): “information as an agglomeration of matter and energy…that it can be encoded or embodied” (p. 60)
- Bawden (2007): “information as embodied, as a self-organizing complex physically present entity” (p. 60)
Again, Budd examines each of these theories and finds them wanting. They simply do not provide us with a decision procedure for determining whether something is or is not an instance of information. He argues that this is sorely needed in information science and promises to suggest a workable definition. Following the thesis of the paper, it’s time for Budd to take a look at truth. But first, some housecleaning
Reference, Meaning, Truth
So, Budd dispenses with the available candidates for definitions of “information” (well, he leaves a few important ones out…more on that later). He is primed to launch into a discussion of truth, but first he has to take a detour through the concepts of reference and meaning…making a few surprising errors along the way.
Budd begins discussing the concept of reference by appealing to Donald Davidson and I’m not too sure about what’s going on here. Not only did Davidson explicitly hold a negative view of reference (cf. “Reality and Reference”, Dialectica, 31: 247-253. 1977) but Budd misquotes Davidson1 and later erroneously credits Davidson with introducing the concept of reference.2 So, what is Budd doing with reference anyway? I can’t tell for sure, but he discusses reference first in terms of aboutness and then in terms of discursive practice…though in both cases he seems to be talking about sense instead of reference. Whatever the case, Budd claims that information seekers have “quite an intellectual and cognitive burden to reach the point where reference is comprehensible” (p. 61) and he drops the subject, turning to an analysis of meaning.
Budd begins his discussion of meaning with the assertion that “reference is an essential element of meaning, but it is not the only one” (p. 61). That’s more like it! On to the discussion of sense! But…Budd never gets around to the sense/reference distinction, and I’m not sure what his angle is. His discussion of meaning touches on indexicality, the analytic/synthetic distinction, speech-acts, semantic prescriptivism, intentionality,3 and rhetoric. I like that he cites Predelli’s work on the context-dependence of semantic utterances as a good direction for information science, insofar as he argues that we need to understand in what ways “contexts are manifest in formal communication” (p. 62). However, the nagging question remains…what does he mean by ‘meaning’? Various technical issues in the philosophy of language are discussed, but meaning itself isn’t really covered at all. I can only guess that his definition of meaning is “that which is understood” which is far from technical.He ends the section by directing readers to Steven Pinker’s 2007 The Stuff of Thought which is a really great and important book on the psychology of language, but not exactly the best starting place for the philosophy of language. Anyway, he then sets up his discussion of truth…
But, is it a fair treatment of truth in the first place? I’m worried that it isn’t. Budd introduces Tarski’s Semantic Theory of truth but quickly dismisses it as “far too limiting for useful application to formal communication of the sort IS is concerned with” (p. 64). He goes on to address the canonical correspondence, coherentist, and pragmatic theories of truth, and finds them all lacking as well (though for rather strange and unconvincing reasons). Ultimately, Budd proposes that we “liberate ourselves from the notion that truth is limited to the meaning of words used in sentences” (p. 65).Yet, just a few pages later he reverses and claims that his definition requires that “any utterance or argument, to be evaluated for its potential truth, must first be meaningful” (p. 68). Granted, he could be saying that the concept of truth applies to both propositional and non-propositional content, but it’s unclear.
Back to information
So, what does Budd propose for a definition of information? Here it goes…
Definition: Information is meaningful communicative action that aims at truth claims and conditions.
Statement of Theory: Information is comprised of those communicative actions (and only those communicative actions) that can be evaluated by a population – defined as the intended or potential hearers of the communication – as meaningful. Meaning is not limited to pure semantics, but includes context and history within evaluation. Further, information is true in that there is warrant for the communicative action, that this action includes no deliberate deception or omission, has inherent evaluative components, provides evidentiary justification, and is fundamental to ethics. (p. 70)
So, there are three necessary and sufficient conditions: meaningful, communicative, truth-directed. Is this a plausible theory? Only if we are willing to accept the following:
- Budd’s theory eliminates environmental information. Consider the dendrochronologist counting tree-rings in the middle of the forest. She determines that a particular oak is 147 years old and writes it down in her log-book. It would seem that she has gathered some information about the tree, though in the absence of a communicative exchange, Budd’s theory would say she has not.
- Budd’s theory eliminates instructional information. For example, a recipe for a cake may include the instruction, “Slowly mix in the flour until the dough forms a ball.” Obviously, you can’t call this instuction “true”, so it fails Budd’s test for information. Indeed, the whole recipe fails the test because it is not aimed at truth claims.
There are other potential counterexamples, but I’ll just list the two. My larger concern is that I just don’t think Budd has established workable accounts of truth or meaning. He claims that meaning includes semantic content, historical context and a “phenomenological element” of intentionality, though he offers little explanation. In light of his discussion of communication, it almost seems as if he is confusing theories of meaning with the related, yet distinct, speech act theories. His discussion of truth is similarly vexing: truth is independent of semantic content and it is neither “naively objectivist” nor “entirely subjective” (p. 71). So what is it? Budd does not tell us, though he offers several competing theories we can reject. Perhaps that’s the takeaway. Budd’s analysis of information takes cues from so many disparate (and sometimes contradictory) areas in philosophy, linguistics, and psychology, that the end result winds up seeming rather confused. I’ll give the benefit of the doubt and assume its just me that’s confused.
I’ll refer to Budd’s theory as the Communicative Theory of Information; I’m sure it will be fleshed out in more detail in the future, but for now we should treat it as just a sketch. But, is it a sketch that can inform library practice? Budd offers the example of information retrieval:
An individual asserts a query…which includes a finite set of elements. Most of the elements are unknown by the individual, but reside within some frames of known and potentially knowable bits. If the individual is seeking information as it is defined here, a search can be constructed that can…result in a set of items that are meaningful and true. That is, the items can be evaluated for meaning and truth. The individual, following the theoretical principles as stated above, has criteria to use in evaluating meaning and truth. Each item retrieved can be evaluated in such a way. That said, the individual is not likely to intuit the assessment mechanisms. In other words, the definition I suggest is usable only inasmuch as it is used by information seekers. (p. 71)
I don’t know what to make of this example, and, damn it, I’m tired of not knowing what is going on in this paper. I can at least make out that the example shows how the Communicative Theory of Information takes information out of the world of data and reduces it to particular discursive practices. But, hold up. Do we really want this? It’s good news for information seekers, but bad news for information systems. Yes, bad news for librarians. Though Budd’s theory may provide insight into search behavior and the social life of information, it is of no help for understanding the systems that collect, analyze, and organize information. Limiting information to Budd’s interpretation of communicative practices makes information too relational and subjective to effectively work with. In sum, we need a different definition of information in library science. I mentioned earlier that Budd left some important definitions of information out of his survey, and I’d like to advocate for one in particular: the Semantic Conception of Information. I’ll get a defense of the Semantic Conception up as soon as possible. Until then, I recommend you take a look at “Meaning, Truth, and Information: Prolegomena to a Theory” and see what you think.
(1) Budd writes:
A truthful sentence is also a meaningful sentence. Davidson (1984), in his early writings, criticized this kind of connection: “any meaning of a sentence is what it refers to, all sentences alike in truth value must be synonymous – an intolerable result” (Davidson, 1984, p. 19).
Davidson originally wrote:
Hence, any two sentences have the same reference if they have the same truth value. And if the meaning of a sentence is what it refers to, all sentences alike in truth value must be synonymous – an intolerable result. (my emphasis)
This is an important distinction. Davidson was criticizing a naive Referential Theory of Meaning. He was not criticizing the connection between truth and meaning, as Budd implies. In point of fact, Davidson’s entire contribution to the philosophy of language was based in his truth-condition of meaning; that to know the meaning of a sentence is to know the circumstances under which that sentence would be true.
(3) Though he seems to be confusing intentionality (the ‘aboutness’ of mental states) with the ordinary language sense of intention (a deliberate determination to act a certain way, and in this case to communicate)