​

Preamble: This post continues my exploration of Patrick Hanks’ Theory of Norms and Exploitations which I had undertaken in my previous post, this time focusing on the notion of meaning potential. Much of the following is based on Chapter 3 “Do Word Meanings Exist?” and parts of Chapter 4 “Prototypes and Norms”, Chapter 5 “Contextual Dependency and Lexical Sets”, and Hanks (The Syntagmatics of Metaphor and Idiom, 2004).

I hope to eventually understand Hank’s approach to metaphor as well, which I have been thinking a lot about as I work through my MA thesis as someone who was taught Lakoff & Johnson’s Conceptual Metaphor Theory as my entry into the cognitive science of language. Therefore, expect Part 3, which should come soon, to be about the wonderfully confusing world of metaphor!

Meaning Potentials and Meaning Events

Usually, when we speak of the impact philosophy has had on linguistics, we often have in mind those philosophers of language within the analytical tradition. These philosophers (e.g., Montague, Russell, early Wittgenstein, Frege) were innovative in their applications of formal logical analysis and reasoning, borrowed from mathematics, to problems of natural language — particularly problems of meaning (i.e., semantics).

It is not often that we see the philosophy of Alfred North Whitehead, a rather different philosopher (even maybe postmodernist) who did not explicitly work on problems of language, bleed into our modern semantic theories. It is not clear to me whether Patrick Hanks, in his Lexical Analysis (2013), was aware of the Whiteheadian influence on his work, but he had doubtlessly inherited it from his predecessors and linguistic influences (John Sinclair, Michael Halliday, and, most tellingly, J.R. Firth). In any case, the quote from Hanks (2013, p. 65) above is an unquestionable derivative from Whitehead’s way of thinking scientifically.

Whitehead’s metaphysics suggests that the fundamental building blocks of reality are events rather than substances. The actual world is made up of events unfolding in the present moment. Events themselves are actualisations of potentials (i.e., possibilities) that are conditioned and constrained by past events which have already occured.

In other words, potentials concretise into events in the actual world, which then pass on to become potentials for future actual events.

Hanks takes this basic Whiteheadian concept and applies it to linguistic semantics. A lexicographer by both training and practice, Hanks suggests that we should not view meanings as something which words have, but which they can potentially actualise. Thus, words do not have meanings. They have meaning potentials.

This is meant similar to how a cannonball atop a tall hill has gravitational potential which can be actualised as kinetic energy or how electrons in a light bulb have electric potential which can be actualised as heat and light energy.

How are meaning potentials actualised? For Hanks, the answer lies in lexical collocation. That is, meaning potentials are realised when a word is fitted among other words into a text. To better understand how this works, however, we need to understand first how meaning potentials are structured.

Meaning Potential of Nouns

According to Hanks, a meaning potential (at least, that of nouns) consists of semantic components. Semantic components are “separate, combinable, exploitable entities” possessed by a word (Hanks, 2013, p. 70). For example, the meaning components of the word bank in its primary sense related to financial institutions can be listed as follows (from Hanks, 2013, pp. 70-71):

• IS AN INSTITUTION
• IS A LARGE BUILDING
• FOR STORAGE
• FOR SAFEKEEPING OF THINGS
• ESPECIALLY FOR SAFEKEEPING OF FINANCE/MONEY
• CARRIES OUT TRANSACTIONS
• CONSISTS OF A STAFF OF PEOPLE

Importantly, none of these above semantic components are necessary conditions to qualify some entity to be referred to as a bank. For example, sperm bank or blood bank are approriate terms despite the fact that they do not safekeep money or carry out transactions in any ordinary sense of the term transaction. As another example, in the sentence He assaulted them in a bank doorway (from Hanks, 2013, p. 69), the semantic component of the bank as an abstract institution is not primarily referenced. Although it is obvious that the assault took place in a building that houses such an institution, it is the building semantic component that is primarily activated.

Thus, Hanks contends that the traditional Aristotelian method of analysing meaning in static terms of necessary and sufficient conditions is obsolete.

Meaning Potential of Verbs

Hanks views the meaning potential of verbs in terms of semantic valency structure, which in turn are expressed in clause structure. This seems commensurate with Robin Fawcett’s views in Cardiff SFL wherein the main verb is treated as a direct element of the clause and other clause elements realise semantic arguments of the verb. There also seems to be compatibility here with Adele Goldberg’s treatment of argument structure constructions in Construction Grammar.

As an example, in Hanks (The Syntagmatics of Metaphor and Idiom, 2004, p. 4), Hanks gives the semantic valence of the verb toast as follows:

 1. [[Person]] toast [[Food = Bread | Nuts]]

This describes the primary meaning of toast as an activity done by a person to a food item (bread or nuts) such that the item is ‘made brown and crisp by exposure to radiant heat’ (2004, p. 4). In other words, to toast bread or to toast nuts.

The verb toast also has a less salient meaning as follows (where “POSDET” stands for “possessive determiner”):

 2. [[Person 1]] toast {[[Person 2] | [Achievement]] | {[POSDET] memory}}

This is the sense of toast when we speak of toasting a significant person, achievement, or person’s memory. It is to ‘celebrate by raising a glass containing alcholic liquor and then drinking some’ (Hanks, 2004, p. 4).

Importantly, this example of toast demonstrates a key tendency: that we can tell a lot about what a verb means by examining the semantic types of its arguments. We disambiguate the first meaning of toast from the second by the type of object it takes — food for the first and person, achievement, or memory for the second.

However, it is also important to acknowledge that we cannot tell all about a verb’s meaning just based on its arguments. Hanks (2013) uses the polysemous English verb check as an example. Considering two senses of check, either as ‘to inspect’ or ‘to cause to stop or slow down’, Hanks (2013, p. 80) shows us the following excerpt from a text in the British National Corpus (BNC):

 3a. He soon returned to the Western Desert, where, between May and September, he was involved in desperate rearguard actions—the battle of Gazala, followed by Alamein in July, when Auchinleck checked Rommel, who was then within striking distance of Alexandria.

It is evident that checked in 3a invokes the ‘to cause to stop or slow down’ meaning, but this information is not inferrable on the basis of the arguments of checked, which are the names of the World War II generals Claude Auchinleck and Erwin Rommel.

Hanks conducted a small experiment in which he presented an edited version of 3a to teenagers and asked them what checked meant in this context. He showed 3b to the teenagers, which was edited to remove miltary-related collocates from the text. It is also important to note that the teenagers were unlikely to be familiar with who the names “Auchinleck” and “Rommel” referred to.

 3b. He soon returned to the desert, where, between May and September, he was involved in actions when Auchinleck checked Rommel.

The teenagers reported confusion as to what checked could mean, despite their familiarity with the word and that some of them were chess players. They reported that the verb “did not seem to fit” (Hanks, 2013, p. 80), though they agreed that Auchinleck and Rommel were probably human beings. One of the teenagers proffered the ‘to inspect’ interpretation as a possible meaning, thinking perhaps that Auchinleck was inspecting Rommel for fleas or smuggled goods.

Hanks thus contends that real-world knowledge regarding Auchinleck and Rommel as warring generals of opposing armies would have been necessary to interpret 3b correctly. In particular, he suggests that the generals’ names would have to be interpreted as metonyms for the armies which they led. Then, it would be straightforward that when one warring army checks its opposing army, here it is meant check as ‘to stop or cause to slow down’ rather than ‘to inspect’.

This metonymic meaning would require two semantic valency structures which are mapped asymmetrically, as in 3c (from Hanks, 2013, p. 81):

 3c. [[Human = Military Commander~1~]] checked [[Human = Military Commander~2~]] = [[Army~1~]] checked [[Army~2~]]

Therefore, much beyond the immediate arguments of a verb are require to discern its meaning. The meaning potential of a verb is actualised not only by what nouns are fitted into its argument structure, but also by what conceptual or mental model the verb itself is being fitted into.

Projecting Meaning Potentials onto Syntax

Consider the following examples from the BNC, as reported by Hanks (2013, p. 138):

 4a. Jani banked £60,000 through successful libel actions against Options magazine and the London Evening Standard.
 4b. She is believed to have *banked £10 million since being booted out of Downing Street two years ago.
 4c. The plane banked, and he pressed his face against the cold window.
 4d. I banked the aircraft steeply and turned.

The way in which Hanks maps semantic structure onto syntax is rather simple. In 4a–4d, three distinct meanings of the verb to bank can be teased out. Similar to the discussion on to toast above, we can disambiguate these meanings of to bank on the basis of its arguments, particular the semantic types of the arguments.

For 4a and 4b, which use to bank in a financial sense, we expect that the subject be of the semantic type [[Human]] and the object of the semantic type [[Money]]. This is shown as a syntactically-annotated valency structure in 5a. Additionally, Hanks provides the corresponding impliciature for 5a in 5b, as what might appear in a dictionary entry for this sense of to bank.

 5a. [Subject [Human]] bank [Object [Money]]
 5b. Implicature: [[Human]] deposits or invests [[Money]] in a bank or other financial institution for safekeeping

For 4c, we find the appropriate semantic valency structure and implication in 6a and 6b respectively. Likewise, 7a and 7b describe 4d.

 6a. [Subject [Aircraft]] bank [NO OBJECT]
 6b. Implicature: (Pilot of) [[Aircraft]] raises one wing higher than the other in order to change direction

 7a. [Subject [Human]] bank [Object [Aircraft]]
 7b. Implicature: [[Human = Pilot]] causes [[Aircraft]] to raise one wing higher than the other in order to change direction

Hanks (2013, p.138) makes a point that the “clause structure of the implicature does not have to be identical to the clause structure of the pattern itself”. By “pattern”, Hanks seems to refer to the semantic valency structure. Hanks’ point, then, is illustrated by the relation between 61 and 6b, wherein the subject of 6a is of the semantic type [[Aircraft]] but the subject of 6b the pilot of the aircraft, and thus technically of the semantic type [[Human]].

One may argue that the subject of 6a being [[Aircraft]] rather than [[Human]] might be a case of metonymy where the aircraft stands for the pilot of the aircraft, similar to the metonymy in 3c, but it is necessary to note that 4c is vague as to the cause of the plane banking. That is, the plane need not have banked as a result of the pilot’s intentional action. Thus, in the implicature of 6b, the role of the pilot is put in parentheses, signifying that the involvement of the pilot is an optional aspect of the implicature, not a necessary one.

The crux of Hanks’ position on the relation between semantic valency and syntactic structure thus seems to be this: the semantic valency structure of a verb specifies at the same time (1) what semantic types are normally expected as arguments by the the verb and (2) what syntactic elements the arguments are normally expected to be realised as (e.g., Subject, Object/Complement, Adjunct, etc.).

Interesting Quotes:

• “There is an interesting and debatable relationship between this point of view [that words have no meaning outside of context] and the principle of compositionality in predicate calculus, introduced by Frege (1884). The principle of compositionality regards words as the bricks used to build propositions. Just as a brick is not a building, so a word in itself is not meaningful. Frege argued that terms expressing concepts are variables that have meaning only when they are put together to form propositions.” (Hanks, 2013, p. 65)

• “There are no necessary conditions for meaning, but some at least must be present if the word is to be used successfully to create a meaningful communicative event. It is an open question how many components constitute a quorum for effective communication.” (Hanks, 2013, p. 71)

• “It is a convenient shorthand to talk about ‘the meanings of words in a dictionary,’ but strictly speaking these are not meanings at all. Rather, they are meaning potentials [emphasis in original] — potential contributions to the meanings of texts and conversations in which the words are used, and activated by the speaker who uses them.” (Hanks, 2013, p. 73)

• “…meanings are best viewed as dynamic events that involve cooperation between speakers and hearers or writers and readers, rather than static, finitely bounded abstract objects.” (Hanks, 2013, p. 73)

• “Outside the context of a meaning event, in which there is participation of utterer and audience, words have meaning potential, not deterministic meaning.” (Hanks, 2013, p. 82)

• “The meaning potential of each word in a text is activated cognitively in the mind of the reader or hearer by other words in the contexts in which it is used. These cognitive components are linked in a network [my emphasis] that is both cognitive for each individual member of a language community and social for the language community as a whole. This network provides the whole semantic base [my emphasis] of the language, with enormous potential for saying new things and relating the unknown to the known.” (Hanks, 2013, p. 82)

• “Viewing meanings as events that take place in the world, in which the participants are utterer (speaker or writer) and audience (hearer or reader), enables us to see more clearly the nature of the lexicon (as a store of shared beliefs or meaning potentials, each of which is associated with one or more phraseological norms)” (Hanks, 2013, p. 87)

• “Each participant draws on his or her mental stock of beliefs about word meaning to construct an interpretation of the meaning events in which he or she participates. A meaning, in this account, is an event involving at least two cognitive participants.” (Hanks, 2013, p. 87)