The Unmasking of Dictionaries by Strong Opinion


I am currently funded by money from Cambridge University Press and Assessment, which also stewards various English dictionaries. All opinions in this post are distinctly mine.

The Unmasking of English Dictionaries (CUP, 2018) is on a mission to change lexicography forever and on the way it tries to insult as many lexicographers as possible. Its author, the linguist R. M. W. Dixon, has a chip on his shoulder. Dictionaries of the English language are all wrong. Their creators misunderstand what a dictionary is for, and are, in general, lazy plagiarists. Surprisingly, the book is not just entertaining, but also makes intriguing suggestions, even though some of the arguments for them have serious gaps.

Dixon repeats again and again that the purpose of dictionary is to “tell you when to use one word rather than another” (p. ix). That is the premise, and on its basis Dixon discusses the shortcomings of dictionaries:

  • They treat words in isolation, rather than contrastive in their semantic field
  • They rely excessively on definitions
  • They neglect to provide grammatical information that is required for correct word use

The examples Dixon gives suggested that there is at least something to his diagnoses. He proposes that the problem be solved by the construction of a new dictionary organised semantic fields. The entries for these fields would compare the usage of the words contained in it, including the grammatical constraints on this usage. For illustration, the book contains a few sketches of such comparative discussions of lexical semantics, e.g. the field including “want”, “wish”, “desire”.

One might, however, wonder about the correctness of Dixon’s premise that the main purpose of dictionaries is to enable the choice between words to use. I don’t think its true, at least descriptively. Dixon takes a distinctly productive task as the purpose for a dictionary, choosing one word to use over others. I expect, however, that much of the use of dictionaries is receptive. My expectation is that dictionaries are most frequently consulted when one is stumped by previously unfamiliar word in a text one tries to comprehend. Of course, that is speculation on my part, but so is Dixon’s claim that the purpose relates to productive use of English. And that leads us to the heart of the problem, Dixon does not sysematically engage with users of dictionaries – others than himself, that is – even though the whole point was to propose a new type of dictionary that is better suited for the needs of its users.

After another swipe against lexicographers as lazy copyists, Dixon proposes the following procedure for producing a dictionary (p. 25-26):

  1. Select sets of related words
  2. Consult corpora to compare and contrast those related words
  3. Work out a conceptual template for the sets of words using “critical notions”
  4. Only at the last step should one compare with other scholars and dictionaries.

Dixon’s proposed procedure does not include users at any point. There are no user studies, not even a step to incorporate informal feedback. The goal is to work “form first principles and with a fresh viewpoint” (p. 25). These first principles might please a linguistic expert such as Dixon, but surely the average dictionary user has different needs and these should be assessed in the process of constructing a dictionary.

The lack of considering actual users and their needs also shows up in another assumption by Dixon, namely that contrastive but relatively abstract outlining of different usage patterns is sufficient to help with the choice between words. Dixon would have dictionaries present everything that is required for choosing between words. That includes a lot of rather abstract linguistic information. The distinction of different types of clauses might quickly overwhelm a learner who just wanted to understand what “hanker” meant in a text they were reading, and while Dixon does envisage the usage of sentential examples, the theory-driven contrasting of words in a semantic set comes first as the organising principle (see p. 227).

I would also like to add that for someone who emphasises “first principles”, Dixon does not spend much time on actually laying out his theoretical framework for lexical semantics. From Dixon’s approach, I would assume that he endorses some sort of lexical relation/frame semantics, but it is not obvious that these approaches correctly reflect word senses as they are cognitively encoded. Surely the first principles for organising a dictionary would be principles that reflect the entries in our mental lexicon? The problem here might be that Dixon does not think highly of many efforts of investigating “the role of language in human cognition” (p. 192). Although my main criticism is that the focus on linguistic “first principles” is to the exclusion of empirically assessing dictionary user needs, it might be noted that even the claimed “first principles” are not exactly fast foundations (using the word “fast” here in the antiquated secondary sense discussed on p. 131-134).

Dixon’s book has great entertainment potential, especially for those of us who enjoy academic philippics. As is common for this text genre, the positive argument reveals holes upon closer expectation. Dixon’s assumption should be considered expert guesses about dictionary use, but guesses they remain. That being said, investigating Dixon’s proposals in actual user studies might be of great interest. The results could show to which extent lexicography really needs to be reborn.

Previous Next
We Know So Little Transformers and the Brain:...