This is a welcome and – as far as I am aware – unique view of plain legal English.
Professor Williams is native English and graduated in Modern Languages before teaching English in Italian universities for forty-four years. His research focused on plain legal English, and he has been for many years the Italian representative of Clarity, an international organisation founded by English lawyers to promote clear legal language and design. This book was his immediate project on retirement.
It was intended to ‘arouse curiosity in readers who may have little or no expertise in plain language, or in legal matters, or in linguistics. The risk is that of disappointing readers who do have expertise in those areas’ (p. ix). Inevitably, a thorough investigation of so profound – and continuing – a change in the style of so many documents, and types of documents, from so many sources, and of so much speech, needs substantial further research. This study could only have been a preliminary guide. I understand it as a call for that research, and it should stimulate it rather than (as Williams fears) disappoint those capable of conducting it. The details, summarised below, are interesting.
I was aware of two limitations, both unavoidable in the circumstances. It makes only passing references to Scottish legal English (which is incomprehensible even to English lawyers); and its contractual examples are restricted to banking and insurance terms, all from the internet. Future research could usefully focus on the language of city and high street lawyers in private practice, the influence of which is much wider. Meanwhile, this introduction to the problems is well written and easy to read. Detailed references, and lists of all the documents in each corpus, are given at the end of each chapter.
Chapter 1: Contextualising plain language (pp. 1–25)
To draw us in, chapter 1 starts with a joke about legalese before settling into the serious business.
First, what constitutes plain language? Williams briefly considers alternative expressions in a selection of European languages and quotes the English ‘definition’ (more accurately ‘description’) adopted by the International Plain Language Federation – which is flawed in my view. He does not adopt it or quote any of the many other descriptions and definitions. This is sensible: precise definition is impossible as the concept is inherently vague and understood differently by different people, and it does not matter that the definition is imprecise because we know well enough for his purposes what he is talking about.
Section 2, headed ‘Theoretical background’, has subsections on deliberate change to language; popularisation (contrasting clarification and simplification); the impetus given by computers and the internet to public access to the law (improving the organisation and design of documents as well as the language); the ethics of plain language; and matching the language to the audience.
Chapter 2: A history of plain language in the legal sphere in the United Kingdom (pp. 26–60)
The introductory section acknowledges that its focus is on England and Wales (still a single jurisdiction despite partial devolution); briefly distinguishes the common law system of the UK (inherited by countries that have been under its control) from the civil law system prevalent in the rest of Europe and similarly exported elsewhere; and considers the mix of languages from which Modern English developed.
Section 2 outlines the influence of Latin and Norman French on the evolution of legal English, with unnecessary archaism encouraged by the doctrine of precedent. This conservatism is particularly unhelpful when applied, as it is, to convoluted style as well as to the meaning that has been attributed to words by judges in contexts that no longer apply. And it has been consolidated by lawyers’ tribalism. (There are, of course, other factors discussed in the works cited – for example, the fear of ambiguity – but they are not important here.) Radical change is now encouraged by the plain language movement within the profession and by commercial pressure from outside – helped, no doubt, by the current tendency towards egalitarianism and a mistrust of experts. (But, as Mattila (Reference Mattila2013) pointed out, this has happened before in various legal languages.) The section ends with a seventeen-item selection of features of legalese.
Section 3 considers objections to legalese by men of letters from Chaucer onwards, including King Edward VI in 1551. Encouragement for change began to appear from senior members of the profession in Victorian times but it was isolated and ineffectual until the 1960s. Williams goes on to describe the main official and unofficial initiatives in the UK in the last fifty years, and it was good to see non-lawyer Martin Cutts given due recognition: his 1993 revision of the Timeshare Act 1992 met the challenge from a sceptical Office of the Parliamentary Counsel (OPC) to produce an effective plain alternative, and in so doing helped change OPC’s culture and improve subsequent legislation. Some space is given to Francis Bennion’s concern that plain language can mislead the public into thinking they understand the law while they overlook the deep meaning of the text.
A short final section suggests that ‘looking at plain language simply as a “movement” is arguably reductive and misleading’ (p. 52) because the word implies grass-roots activism. I am not convinced it implies this, but does it matter?
Chapter 3: The language of legislation (pp. 61–92)
Section 1, about the available corpora, functions as an introduction to this and the two following chapters. It has a useful list of relevant corpora but regrets that those ‘available specifically on legal English in the UK all focus on the language – spoken and written – of the courts’ (p. 62). So Williams has compiled small corpora of his own and has (successfully) aimed to present his results non-technically to help readers without linguistic knowledge.
Section 2 describes the corpora he has compiled to compare 27 public general Acts passed in 1970 with 38 passed between late 2018 and early 2020 (which for simplicity I’ll refer to just as ‘2020’), and mentions the Sketch Engine tool he used to help manage the data. Section 2.2 lists the seven features chosen, not to categorise text as plain or traditional (which would be impracticable, controversial and unreliable) but ‘that could provide evidence as to whether plain language techniques have been introduced’ (p. 66).
The results are analysed in section 3, each feature in a separate subsection. (a) Sentence length is of course ‘a very crude yardstick’ but ‘as a general rule' a useful guide (p. 67). It is also laborious to calculate: counting full stops is very far from good enough. Williams’ calculation gives a reduction in average sentence length in the later corpus from 77.1 words to 47.7. (b) The reduction of unfamiliar (and often otiose) pro-forms like aforesaid and the foregoing in the same period is more dramatic: the never-useful the said, for example, drops from 381 occurrences to none. (c) Unfamiliar pronominal adverbs such as hereby and whereof have disappeared entirely. Thereof, with 149 instances, was the most prolific of the 17 words searched in the 1970 corpus; all were reduced to 0 in 2020. This is ‘one of the clearest indications that plain language principles have been taken to heart’ by OPC (p. 72). (d) Gender bias has also fallen dramatically, notably by avoiding pronouns. Williams mentions only in passing Xanthaki’s unconvincing (Reference Xanthaki and Flückinger2019) distinction between ‘neutrality’, which ‘can [my emphasis] be viewed as an expression of feminism’ and non-binary ‘inclusivity’. (I was persuaded that neutrality is a plain-language point when a judge said that his juries never selected forewomen when he instructed them to appoint a foreman.) (e) In an effort to avoid stigmatising the proper use of passive verbs, only transitive finite verbal constructions were considered in comparing the 1970 and 2020 active–passive ratios. And because of the need to consider each such verb in context only 20,000 words were selected for each year. On this basis the percentage of passives fell from 54.7 to 32, reflecting OPC policy to prefer active verbs (or a succinct imperative) where it improves readability. (f) Between 1970 and 2020 the legislative use of shall decreased from 13,435 instances per million words to 70 per million. OPC guidance is to avoid it unless there is good reason to use it (for instance in ‘text to be inserted into an Act that already uses it’, p. 84).
Chapter 4: The language of court judgments (pp. 91–135)
Section 1 distinguishes (for non-lawyers) court judgments from law reports, and (for non-linguists) Crump’s (Reference Crump2002) persuasive texts from preservation texts. (If it matters, I disagree that while judgments are ‘arguably a combination of both’, ‘legislative texts come firmly within the latter category’ (p. 94); like most other legal texts, both types are intended to influence conduct and perhaps attitudes, as well as to record.) There follows a discussion of the free style of UK judgments, in contrast to the more rigid and abstract judgments in France and Germany (Mattila Reference Mattila2011: 97–104). But the engaging language of some judges, in particular Lord Denning, is not used by those many more of whose style ‘there has been some robust criticism’ (p. 95). Williams also records an emerging style of judgments specifically addressed to children and others, notably in sensitive family cases.
Section 2 introduces his 1970–1 and 2020 selections from the Court of Appeal, each of 100,000 words from the civil division and 100,000 from the criminal. These ‘judgments corpora’ are reduced to rather smaller ‘judgespeak corpora’ by omitting words, such as quotations, over which the judge had no control. Williams explains the need to use a different set of features from those in chapter 3, and the difficulty of selecting them.
Section 3 analyses the results. (a) Average sentence length decreased from 27 to 25 words in the criminal judgments but increased from 25 to 29 in the civil cases. These figures are low by traditional legal standards and not far above common plain language recommendations, and I wonder to what extent the increase in the civil judgments can be attributed to Lord Denning’s retirement. Renowned for exceptionally short sentences and the clarity of his language in general, he was the lead or only author of two of the ten earlier judgments. (b) The length of judgment was calculated from 100 early and 100 late cases in each division, taken over a longer spread of years. The detailed results are too intricate to summarise, but civil judgments tended to be much longer than criminal judgments and increased by a higher percentage over the period. Various possible reasons are suggested, with the rider that more research is needed to consider each judgment individually. (c) No unfamiliar pro-forms were found in either period except when they were part of quoted text. (d) By contrast, judges' own words included 44 unfamiliar pronominal adverbs in the early corpus and 30 in the later one. But they were relatively rare and the one quoted sounded to me quite reasonable. (e) Passive verb use remained surprisingly high at about 30 per cent over the period, although we do not know how many occurrences were appropriate. (f) The frequency of the 26 selected Latinisms combined dropped from 91 instances to 35. Again, the influence of one judge may have been influential: Lord Woolf’s dramatic plaining of the civil court rules replaced many Latinisms. Or it may just be that fewer judges had been to public school. (g) Against the trend towards linguistic informality, the incidence of first-person pronouns and adjectives has fallen substantially, particularly in the use of I. Williams suggests that a desire for objectivity may have outweighed the general trend towards a more personal tone but he calls for further research. (h) The use of titular nouns and other forms of courtesy has fallen. Mr remains far more frequent than the female equivalents, which have risen slightly, with Ms replacing Mrs but the use of Miss increasing. Are there now more female litigants than before? The references to fellow judges as learned (often when criticising them) dropped from 214 to 5.
The results are summarised in section 4. On the basis of these data, plain language principles seem to have had relatively little impact on judgments, and much less than on legislation. Williams suggests that judges be given guidelines, as parliamentary counsel are, and a word limit for judgments is suggested to rein in prolixity.
Chapter 5: The language of online terms and conditions (pp. 136–77)
Section 1 outlines the categories of lawyers, recent changes in working practice, and how the corpora were selected when contracts (and most other documents written by UK lawyers) are private, and drafted by so many different lawyers for different purposes. Those available for comparison are much more limited, except in length. Williams justifiably uses the doublet terms and conditions (conveniently abbreviated to T&C) because it is entrenched in the language; in fact, although clumsy, it disambiguates: terms alone could refer to the whole set or to individual terms within it, and condition has a specialist legal meaning. He concedes that online terms are not an ideal example of contracts because they are not negotiable, but that does not matter for his purpose; on the contrary, they highlight the need for plain language to balance bargaining power. There were no online terms in 1970 but he found that many online T&Cs are clearer than traditional T&Cs, although still too verbose to read (and no doubt longer than ever). A few firms help customers by improved design, including non-verbal aids.
Section 2 gives details of consumer protection, statutory and otherwise.
Section 3 first describes the division of the 109-document corpus into B2B (business to business), subdivided into banking and non-banking; B2G (business to government); and B2C (business to consumer), again subdivided into banking and non-banking. All belong to the period 2017–21. The choice of different plain language criteria is explained.
Section 4 analyses the results. (a) Sentence length varies between 30.6 words per sentence in the B2C banking texts and 56.6 words per sentence in the B2G texts. B2C texts are on average three-quarters of the length of B2B. B2G sentences are the longest. So an effort is being made for consumers. This suggests to me that businesses have not yet realised that everyone benefits from plain language – themselves included. (b) The texts are typically so long and boring that no customer can be expected to read them before ‘agreeing’ the terms. Williams’ extreme example is Paypal’s, according to Which? (Parris Reference Parris2012) containing more words than Hamlet’s 32,197. (But Paypal seem to have taken this criticism to heart, their late-2022 version having been reduced to 21,942 words.) (c) Unfamiliar pro-forms are greatly reduced, especially in B2C texts. (d) But unfamiliar pronominal adverbs cling to life, especially B2B. (e) Gender language remains mostly male-based. (f) The greater preference for the active voice in B2C than in B2B texts again shows that business lawyers can write more readably when they want to. (g) But their use of shall every 153 words suggests that they do not often want to (even if one accepts Kenneth Adams’ view (Reference Adams2017 and elsewhere) that its use is worthwhile in some contexts). (h) First- and second-person pronouns and adjectives are helpful in unnegotiable contracts imposed on the reader, and are widely used in the B2C texts. (i) Contractions are commonly used in this corpus, even in B2B contracts although with rather more B2C. There is no indication of the frequency with which they are avoided. Williams does note the seeming consensus that they are still inappropriate in the most formal documents, although ‘rapidly becoming the norm in persuasive legal writing’ (p. 167) (and, I suspect, more generally).
The conclusion in section 5 is unsurprising: styles range from traditional to very informal.
Chapter 6: Visualising the future (pp. 178–97)
Section 1 summarises the three previous chapters and adds that ‘the impact of plain language is clearly visible’ in other areas (p. 178).
Section 2 considers design (the arrangement and appearance of documents) and visualisation (replacing text with pictures, charts and other images). Both are recent innovations in legal texts, although visuals have clarified the legally binding Highway Code since 1931. Visuals have begun to creep into contracts and ‘many major law firms have legal design teams’ (p. 187).
Section 3 looks at the effect of computerisation – and COVID – on legal practice. Although the internet is encouraging lawyers to be user-friendly, it is only beginning to erode the use of legalese.
I did not find section 4’s review of ‘the ebb and flow of plain language’ particularly helpful (pp. 189–93). This is partly because of the limitations, acknowledged by Williams, of his research criteria and tools on this point. But it is also restricted to the flow of the last fifty years without the perspective of the longer-wavelength ebbs and flows since Hammurabi’s Babylonian Code; the longer view might have slightly dampened the optimism with which he concludes. However, that does not detract from his conclusion that plain language is slowly becoming the norm.
Reviewer's conclusion
Since the 1960s there has been much encouragement for lawyers to abandon legalese. It has come from the public (as individual clients and through consumer organisations); national and European governments (sometimes by legislation); the Law Society and the Bar Council; some judges; professional literature (which seems to go largely unread); and a small minority of campaigning lawyers. As a result, the old myth that legalese is necessary has been largely dispelled. It is disappointing that so much of it is still produced. Why is that?
I recommend this book in particular to anyone interested in picking up the challenge of further research. Wills might be a useful source: they are very common; written by a wide range of solicitors and some barristers; often relatively short; and notorious for their legalese; and they become publicly available soon after the testator’s death. Other possibilities are the effect (if any) of plain language on the way law students are taught to write and speak and the extent to which any beneficial effect survives their employers’ influence. Some plain language proponents give in-house training: does it work?
