Monday, July 06, 2015

Messing with LCN DNA

The delights of Bayesian probability reasoning are sufficient to draw from me another case comment!

Our Court of Appeal, in Manoharan v R [2015] NZCA 237 (11 June 2015), has said that a likelihood ratio of 20 is “not strong” [52].

This was an LCN DNA analysis and obviously compared to the usual DNA results of likelihood ratios in the many millions, 20 is not high. But in the context of a case it can be, as can be seen from the results of applying Bayes’ Theorem (this was a single-issue case involving independent items of evidence).

I should emphasise that the single-issue nature of this case ("who did it?", an actus reus issue) simplifies the use of Bayes' theorem. Where several elements of the offence charged are in issue, the theorem must be applied to each separately, and only evidence relevant to that issue is used in each application. This prevents, for example, a huge scientific LR on the issue of identity from swamping the issue of intention. For each issue the "probability of guilt" means the probability that that issue is proved to the standard required for guilt to be established.

The likelihood ratio used by the scientists in declaring a match is not necessarily the same as the likelihood ratio for the match evidence used by the fact-finder in the trial. This is because the defence may have an innocent explanation for the match which brings the denominator of the likelihood ratio (the probability of getting the match on the assumption that the defendant is innocent) close to the value of the numerator. Then, the probative value of the test result for the prosecutor would be very reduced or even extinguished.

The scientific LR of 20 to 1 in favour of a match is used to calculate the numerator of the LR at trial: the probability of getting the evidence of a match on the assumption that the defendant is guilty. Often a scientific LR for DNA is in the millions, resulting in a trial LR numerator of approximately 1. But here the numerator will be much less: a ratio of 20 to 1 is a probability in favour of 20/21 = 0.95.

We can compare what the probability of guilt would have been if the scientific LR had been in the millions and the trial LR numerator approximately 1, with what the probability of guilt would be under the revised scientific LR and corresponding new trial LR, for priors (that is, the ratio of probability of guilt to probability of innocence, based on all the other evidence in the case, for this single-issue case) of various levels. (For multi-issue cases the priors will be assessed separately for each issue.)

Fortunately this comes down to a simple relationship: the new probability of guilt is 0.95 times the old probability of guilt. The assumption is that with an LR in the millions, the probability of guilt is 1, whereas with a reduced LR of 20 the probability of guilt is 0.95, the priors being the same in each alternative.

This means that the scientific match evidence isn't as probative as it was when the scientific LR was in the millions, but the reduction is fairly small - although it could make the difference between proof beyond reasonable doubt and acquittal:
  • If the old probability of guilt, on all the evidence at trial, had been 0.95, the reduced scientific LR decreases the final new probability of guilt to 0.90.
  • If the old probability of guilt was higher, say 0.98, the reduced scientific LR reduces that to 0.93.
  • And, if the old probability of guilt was 0.90, the reduced scientific LR decreases that to 0.86.
Another perspective on the size of a LR of 20 is obtained from considering a case where the evidence is not given, compared to one where it is.
  • If, without the evidence, the probability of guilt is 0.90, then with the evidence that probability increases to 0.99.
  • If, without the evidence, the probability of guilt is 0.60, then with the evidence that probability increases to 0.96.
  • And, if, without the evidence, the probability of guilt is 0.30, then with the evidence that probability increases to 0.89.
These latter 3 examples show that evidence with LR = 20 can have strong probative value, and the first 3 examples show that this value is less than it can be when the LR is much higher.

Which perspective was appropriate in this case? The Court was anxious to remove what it called the CSI effect (the jury might wrongly assume that a DNA match was conclusive evidence of guilt), and this is supported by the first perspective. The danger was that the second, and more relevant, perspective would be overlooked.

Perhaps the Court under-emphasised the strength of the DNA evidence in this case. It approved defence counsel’s submission that the jury should have been told that in New Zealand it was likely that there are 200,000 people who would have the same profile. The fallacy is to include people who had no opportunity to commit the crime.

Saturday, July 04, 2015

Search seminar

Notes for the talk on the law of search that I gave at the Auckland District Court on 1 July are available here. They are complementary to the more detailed paper that has been distributed to lawyers.

Sunday, May 31, 2015

Proportionality and the Rule of Law

As you know – although I’m sure you wouldn’t admit it in front of your drinking buddies – there are few more enjoyable ways of spending a wet Queen’s Birthday holiday weekend than by struggling to understand a book written in technical language that you feel you should understand.

So it is with Proportionality and the Rule of Law: Rights, Justification, Reasoning (CUP, 2014), a collection of essays by 18 contributors, edited by Grant Huscroft, Bradley W Miller, and Grégoire Webber.

I can only speak of the Introduction, as the book is rather expensive, and even the ebook seems over-priced. At least Amazon gives us a free sample, which includes the very excellent (as opposed to sort-of excellent?) Introduction by the editors.

So my little game, if you think of it like that, is to translate into ordinary lawyers’ English the technical language of the Introduction. But beware of the risk that I do this with a clarity born of misunderstanding.

“Proportionality” has a range of meanings and can refer to a method or to a goal of decision-making. I wouldn’t want to confuse it with other kinds of decision-making, such as logic, formalism (the application of rules to facts), morality (what would be the morally right decision), pragmatism (what result would work), although there can be some overlap.

Balancing of competing values is a proportionality method of decision-making. So is the rather different ends/means balancing, but this can be seen as a method or as a goal. Using reasonableness to limit what is acceptable is also a proportionality method. Sometimes proportionality endangers rights, in balancing them against other values, and sometimes it compromises moral values, where what is right yields to a greater right or some greater interest.

What happens to rights in proportionality reasoning can vary. Rights are not necessarily eroded in the balancing process, which will usually recognise their enhanced weight by virtue of their status as rights, but if an issue of limitation is being considered then there is a risk of erosion if proportionality requires that.

Proportionality can require recognition of the autonomy and dignity of the person, and this may guide the interpretation of legislation. Legislators, however, may have a greater awareness of rights and social interests than do courts, so executive decisions should be judged by their method rather than their outcome. This concern would limit the role of proportionality reasoning. Indeed, it is arguable (although I am not convinced by this) that proportionality is too abstract a method to be of use to judges.

You could say that morality is important and that proportionality reasoning is not a complete method for judicial decision-making. There are risks attending proportionality reasoning: irrelevancies may be taken into account, things that are doubtful may be treated as certainties, a judge may yield too much to extraneous determinations, aspects of the public good may be ignored, and a judge may have resort to a personal political philosophy.

In their conclusion to the Introduction the editors ask some pertinent questions, which the essays apparently leave the reader to consider. I put these in my own words, sacrificing the subtleties. Does proportionality erode rights? What about absolute rights? Should judges take more account of the reasons that motivate enactments? Should legislators, rather than the courts, use proportionality reasoning? Where proportionality reasoning includes morality, does it prefer some moral theories over others? How should the dangers of proportionality reasoning by courts be overcome?

Well, it’s a shame that the book is so expensive. Too expensive to read. Hopefully my clumsy summary of its Introduction will make some of the essayists’ interesting ideas more accessible.

Saturday, May 09, 2015

I doubt, therefore I am, but what are you?

In Seeing Things as They Are (OUP, 2015) John R Searle gives idealism a long-deserved slap. “There is something tragic about the massive waste of time involved in the whole tradition of idealism.” (P 93, footnote 10, if my Kindle app pagination is accurate.)

Idealism is that philosophy which claims that the only things we have perceptual access to are our own subjective experiences: all we can ever perceive are our own subjective impressions and ideas (Descartes, Berkeley, Hume), we can never have knowledge of things in themselves (Kant), we can only perceive sense data (Ayer). Searle’s position is that idealism leaves us with essentially an unbelievable conception of our relation to the world (p 231).

I am not able to review the book, [Update: here is a review by Josh Armstrong in the LA Review of Books.] but you may wish to view this YouTube clip of a seminar conducted by Searle which substantially overlaps the subject-matter of this book and gives a sense of the technical language generated by philosophical contemplation of perception.

Searle makes an interesting observation about El Greco and whether the painter had defective vision (p 141):
“The hypothesis ... that he painted distorted figures because a normal stimulus looks distorted to him makes no sense, because if he is reproducing on the canvas what produces distortions in him, then he will simply reproduce what looks normal to the rest of us.”

This has implications not mentioned by Searle but which will occur to lawyers. Would El Greco have described in words an obviously distorted image? Are errors in one mode of perception only apparent to other people when translated into a different mode of communication? If a witness describes what was seen, will that description necessarily correspond to the witness’s visual perception? How should a verbal description of what was seen be checked?

Judicial accounts of how facts are determined give no assurance of their correspondence with reality. As EW Thomas observes in The Judicial Process (CUP, 2005) at p 321, “The facts are the fount of individual justice” but there is scope for improvement in the ways they are determined. For example, there is too much weight placed on the demeanour of witnesses (324), and truth, as far as the system will permit “can be gleaned from a close reading of the contemporaneous documentation, if any, or an analysis of the probabilities intrinsic to the circumstances and about which there may be little or no dispute” (325).

As a senior appellate judge, Thomas cautions that
“what judges must not do is fill an unresolvable gap with a judicial ‘hunch’. To do so is to succumb in part to what I have perhaps unkindly labelled the ‘God Syndrome’. Anecdotal evidence suggests that the God Syndrome settles on some judges shortly after their appointment to the Bench ... [and] many appellate judgments would be edified if judges at that level did not show an unhealthy preparedness to adopt a version of the facts which cannot be found in the [trial] judge’s findings of fact or in the transcript of the evidence itself. ... The God Syndrome does not strike at first instance only.” (326)

The resort to assessment of probabilities to assist in determining facts is also referred to by Richard A Posner in How Judges Think (Harvard UP, 2008). He uses (65) Bayesian decision theory to illustrate how, before a witness even testifies, a judge will have formed an estimate that the testimony will be truthful, based on experience with witnesses in similar cases (including when the judge was a lawyer), on a general sense of the honesty of the class of persons to which the witness belongs, or even the way in which the witness enters court and approaches the witness box. It would, says Posner (67), be irrational for judges to purge themselves of this way of thinking.

And the sneakiness of some appellate judges does not escape Posner’s comment (144):
“   Appellate judges in our system often can conceal the role of personal preferences in their decisions by stating the facts selectively, so that the outcome seems to follow from them inevitably, or by taking liberties with precedents.”

(I mention in passing – just to show that some judges do read each other’s books - that at 261 footnote 63 Posner cites Thomas’s book.)

Posner had also discussed the difficulties of ascertaining, from evidence given in the courtroom, the reality of what happened, in The Problems of Jurisprudence (Harvard UP, 1990), particularly at 203-219. He adds (217):
“The celebration by lawyers and judges of the “fairness” of a system in which it is thought better to acquit ten guilty defendants than to convict one innocent defendant is an attempt to put a good face on what is actually a confession of systemic ineptitude in deciding questions of guilt and innocence.”

Ah yes, there’s nothing like a little philosophy to make you have doubts about everything (except your existence).

Friday, April 17, 2015

Book review: Final Judgment – The Last Law Lords and the Supreme Court by Alan Paterson (2013)

Every barrister should read this book. Even if you don’t particularly care about the goings-on in the Supreme Court of the United Kingdom, there are things here about advocacy and the politics of multi-judge courts that make fascinating reading.

The overall quality of Professor Paterson's writing is so good that we should forgive his solitary and incorrect mention of New Zealand. It seems his view of the world is not this one:

There are many reviews of this work online, some of which summarise it in detail. So I will just mention some points and leave you to get your own copy.

Oral advocacy is quite a different exercise from written advocacy. Quoting Michael Beloff QC in Chapter 2:

“One wants to reserve something quite deliberately for oral advocacy— as it were to take a forensic punch, to start off with something that captures their imagination immediately.”

Paterson notes that an appeal before the final court has involved a conversational style of advocacy:

“[Metaphors] ranged from ‘an academic seminar’ or Oxbridge tutorial, to ‘an informed dialogue’, and ‘a dialectic between Bench and Bar’, which resembles nothing so much as a ‘conversation between gentlemen on a subject of mutual interest’. [footnotes omitted]

And one needs to be sensitive to when the court has heard enough:

“Lord Bingham had a way of saying ‘Yes’ which would quicken and multiply if counsel failed to take the hint.”

Questions from the judges can’t be ignored:

“Whilst the judges in the final court can use the dialogue to constrain counsel’s arguments as we have seen, they also use the dialogue to clarify what counsel are arguing, to test counsel’s arguments, and to put their own theories of the case to counsel.”


“ … in general the least helpful thing that counsel can do is to decline to engage in the dialogue. This as Lord Bingham observed, is almost a golden rule of appellate advocacy and reinforces the point that it is about dialogue rather than sequential monologues.”

An anonymised leading counsel is reported as saying:

“A fluent and compelling response to an adverse judicial intervention is the holy grail of oral advocacy.”

And further,

“However clever the Lords are they’re not computers, they’re human beings and you’ve got to make them want to decide in your favour, and that’s what advocacy means, it’s working out a way of making them feel comfortable coming with you.”

Courageous advocacy means having the courage to rely on your strongest point, and not needing to refer to lesser ones. Sir Patrick Hastings KC is quoted:

“The ability to pick out the one real point of a case is not by itself enough; it is the courage required to seize upon that point to the exclusion of all others that is of real importance.”

That’s enough to give you a sense of this topic, which Paterson treats in considerably more detail.

The book concentrates on the various lines of communication that influence judges: with counsel, with other judges on the same bench, with other courts locally and internationally, with parliament and the executive, with law reform bodies. The analysis of statistical information on outcomes is as detailed as it could possibly be, and is a model for the sort of analysis that could be applied to any multi-judge court.

The discussion of how judges decide cases is fascinating too. Meetings before and after oral argument, decisions on who should write the lead judgment, discussions during the writing process (including with the judicial assistants), exchanges of drafts, types of concurrence and the value of dissents are all examined with the benefit of the author’s extensive interviews with the Law Lords and, later, the Justices, as well as with senior counsel. We learn about different types of judicial personality (including observations on Lord Diplock that force one to conclude he was an arse) and the tendency of judges to reflect ordinary group decision behaviour by wanting to stick with conclusions they have expressed at an early stage. It is common for people not to know their own thoughts until they express them in words, and meetings at an early stage may lead to premature conclusions and fixed views. The early drafts of a judgment may force a judge to recognise an error in thinking.

Lord Hoffman is quoted on this:

“ … I must have written about eight or nine drafts of a supporting judgment. I was not satisfied that I’d got it right in any of these drafts and I’d tried this way and that way and eventually it seemed to me the reason why I wasn’t getting it right was because I was wrong. So I changed sides at that point and it went 3: 2 the other way.”

An illustration of interactions and changes of position is given in discussing R v Waya [2012] UKSC 51 (which I have commented on here on 2 January 2013). Paterson describes what went on:

“Since they could not agree on the outcome of the case, it was re-heard before an expanded panel of nine in March 2012. After the first conference (really the second conference [ie the meeting that occurs immediately after oral argument]) the Justices were once again unable to produce a majority position. Lord Phillips suggested that they put their thoughts on paper and eventually a majority position emerged. Lord Reed, who had come in for the second hearing thought that the confiscation order should be set at zero. He laboured long and assiduously to produce his dissent and circulated it before the majority. When the latter came, their position had changed to take account of his dissent. Lord Reed then agreed on a joint judgment with Lord Phillips and withdrew the bulk of his dissent. In all it took 288 days from the second hearing to the final judgment. In part this was a product of the team-working of today’s Supreme Court which eventually produced a conjoined majority and a conjoined minority judgment. However, if the same case had arisen in Bingham’s time it might well have been a case where he approved of a single judgment in order to provide guidance to the lower courts.”

But you mustn’t let me breach copyright by going on like this. Buy the book and read it; if you’re a barrister I doubt you will be able to put it down.

Tuesday, January 20, 2015

Book Review: The Singular Universe and the Reality of Time, by Unger and Smolin

The Singular Universe and the Reality of Time by Roberto Mangabeira Unger and Lee Smolin, CUP, 2015 brings to our attention a likely explanation for the lack of progress in cosmology since the 1970s. The book will be of interest to some lawyers (Unger is, among other things, a legal theorist) for the light it may shed by analogy on some aspects of legal theory.

This is really two books under one cover, both advancing the same general argument, with Smolin addressing some scientific topics at a level of detail, but not at undue length, that is beyond what general readers would understand. The first part of the book is by Roberto Mangabeira Unger, and the second, from page 349 (if my Kindle app is accurate) by Smolin.

Both authors wrote the introductory remarks, from p x to p xx.

“Cosmology is the study of the universe as a whole, beyond which, for science, there lies nothing.” (p xx).

There are “three big questions about cosmology”, stated by Smolin (401) as:

“What happened at very early times, closer to the initial singularity? What will happen to our universe in the far future? What is there, very far away from us, outside our cosmological horizon?”

Cosmology can only answer these by way of being a science. A science is (501)

“... only about what can be conclusively established on the basis of rational argument from public evidence.” [footnoting Smolin, The Trouble With Physics, 2006, a book that I can recommend]

Incidentally, and I digress for a moment: in The Trouble With Physics Smolin gives us (293) Feyerabend’s amusing observation that he could win any argument in philosophy simply by using skills he had acquired as an actor. “This made him wonder whether academic success had any rational basis.” Drama-queen teachers take note!

No progress has been made on important cosmological questions, and Unger and Smolin argue that this is because scientists have used incorrect assumptions. Questions include: why, when important equations have many solutions, is only one accepted as correct? Why are there three spatial dimensions? Why do the so-called universal constants have the values they do? Why is there life? How did the universe come to what appears to be thermal equilibrium at a single temperature? Why do we assume that the laws of physics apply throughout the universe (including the part of the universe we cannot yet detect), and in an unchanging way?

Most of us will remember wondering why Einstein started his special theory of relativity by assuming that the speed of light in a vacuum is constant and cannot be exceeded. True, experiments suggested that that was so, but should those results be applied everywhere and always? His other basic assumption was that the laws of physics apply uniformly for all frames of reference. Do they, and why?  And his removing of time from its role as a universal and uniform measurement may well produce results that appear to be correct in the part of the universe we can currently detect (even if only by measurement to a level of precision only expensively achieved), but everywhere and always?

Unger and Smolin reinstate time as the absolute backdrop to physical events (52), not an accessory to space (53), they assert it is real (354) and it enables recognition of what they assert will be found to be the evolution of physical laws (357). It also gives meaning to causation: if time is not real, causality (the influence that a state of affairs exercises over what follows it) cannot be real (35). However, the uneasy reader might say, it is not explained how this absolute time is to be measured, and how the measuring standard might change as the physical laws evolve. Still, the point seems to be that time does not vary within a frame of reference, or between frames of reference, or historically compared to any measurement that is known. Where this leaves Einstein is not made clear (at least to me, other people are probably not troubled at all), unless the modification is confined to the unknown part of the universe.

Three central ideas – or perhaps we could say assertions or axioms - are developed in the book (I summarise from pp 5-16): the solitary existence of the universe (there are not, and have never been and will never be, mulitple universes); time is inclusively real (nothing is outside time, everything changes sooner or later); and mathematics is not a substitute for reality, it should be seen as representing a world eviscerated of time and phenomenal particularity (mathematical relations are timeless and of a general character, and indeed are useful to a point, but they do not model a universe in which time is real).

At the very least this book illustrates what can be done when a discipline reaches a dead-end. Fundamental assumptions are re-examined and changed.

One of the joys of looking at the problems that have arisen in another discipline is that ideas applicable to law may be brought to mind. Even now, while I half-listen to New Zealand playing Sri Lanka in a one-dayer at Nelson.

Is there an unknown law that has yet to be perceived?
Obviously, yes. All the things that have yet to be invented are potential subjects for new law, and new kinds of laws. It is impossible to imagine what they may be, but we can say that there will be laws that will apply to them. Whether those laws will be recognisable by reference to our current criteria, or whether new criteria for recognising law will evolve, is not known. But we can say that there is no reason to keep our definition of law constant so that it is temporally limited to known conditions.

Does the law apply everywhere?
Courts have had to decide whether constitutional protections apply extra-territorially (for example see the case discussed here on 13 June 2008). There is increasingly a tendency for the law of one territory to be given extra-territorial application, at least as far as the courts of the domestic territory are concerned. And within a territory it seems there is no room for exceptions to the obligation to obey the law; exceptions are within the law, not extraneous to it. The law applies under the surface of the earth and above it, extending into space. Rights may be given over territory on the moon, and beyond. The extent of the application of the law may only be limited by the human imagination. This means it may apply in environments that are little known, raising the question of the extent to which the conditions for application of law are part of the definition of law.

What is progress in law?
Social norms do not have to be laws (see Gardner, Law as a Leap of Faith, reviewed here on 6 July 2013). If procedures are applied to norms, sufficient to make them recognisable as laws, then that is what they are. But some laws are rights, accepted as somehow being of a higher kind than other laws. If that has happened to a norm, the progress is from a non-legally enforceable social convention, to law, to higher law. Increasing recognition of rights suggests increasing opportunity for progress. Law is not static in this regard. Whether rights will continue to increase, or will reduce, is not known, but there is no reason to think that present trends will always apply. To what extent can we expect to be able to prevent, or encourage, change?

Is it useful to ask why a law exists?
People don’t need laws: it is possible to imagine a community that functions without a legal system (Gardner, above, pp 296-301, suggesting that the functions of law can be served by conventions like morality, etiquette, games, and traditions, and law may have different importance at different times in history, and may one day be lost and forgotten).

So what is law for? Convenience, obviously, in settling disputes, minimising conflict, and preventing or making-good harm. Those functions can be achieved by means other than law, so they do not assist in defining what law is. Law is not defined by its use, but by the process by which it comes into existence.

What is the relationship between a law and the environment in which it applies?
The application of a law is different from its definition as law. A law may be a law although the circumstances in which it could apply have never arisen. For example, laws made to apply in the event of natural disasters or epidemics are still laws pending those catastrophes. The law is not imaginary, although its application is. One of the main aims of criminal law is deterrence, and it is made in the hope – optimistic though it may be – that it will never need to be applied.

The environment in which a law applies may change dramatically, while the law stays the same. But this stasis is not essential: the law may be developed, by conventional techniques of legal reasoning, to meet new demands. To what extent is it then the same law as before? Laws, once created, may evolve with the environments in which they are applied.

What is the relationship between law and reality?
Just as in physics mathematics should not be confused with the reality which it is used to explore, so too in law the techniques of legal reasoning should not be confused with the reality to which law has to be applied. Conventional legal reasoning – the techniques of statutory interpretation, the methods of arguing about case law – will not necessarily produce the correct application of law to a particular problem. The life of the law, as has been said, is not logic but experience. Logic, in law, is subject to correction on policy grounds.

To what extent should legal concepts be defined?
The fight in physics is against uncertainty, whereas in law uncertainty can be an advantage. The circumstances in which a law may have to be applied cannot necessarily be wholly anticipated when the law is made, and some judicial creativity in its application should be allowed for, if the law is to remain useful. It is not unusual to find a statute on a subject which itself is not defined, as for example where the Search and Surveillance Act 2012 [NZ] does not define search.

It may not be useful to define concepts when the law’s purpose may have to be applied in circumstances which, at present, are not clearly perceived. We may know what we want, without necessarily knowing when we may want it. Should this desire be expressed as a law?

What are the premises of the rule of law, and should they be the same everywhere?
An idea that has emerged relatively recently is that a fundamental requirement of law is that it should be identifiable, ascertainable, equal in its application, accessible, and, to an extent that is a matter of some debate, consistent with the requirements of a fair trial and other fundamental rights. The debate about this latter point is not about whether trials should be fair, etc, but about whether this requirement of fairness and conformity with fundamental rights is part of the rule of law.

It seems uncontroversial that the law should be ascertainable and equal in its application, but this is not necessarily universally accepted. There has recently evolved the closed-material form of trial, where the defendant is not told what some of the prosecutor’s evidence is. Public interest in security is said to outweigh the defendant’s right to confront and cross-examine opposing witnesses. See further the discussion here on 18 April 2014.

Monday, November 17, 2014

Book review: The Sense of Style - The Thinking Person's Guide to Writing in the 21st Century, by Steven Pinker

I enjoyed most of Steven Pinker’s, The Sense of Style – The Thinking Person’s Guide to Writing in the 21st Century.

My computer puts a green wavy line under the second definite article in that title, apparently because it doesn’t like the capitalisation. Just illustrating that we can all find something to argue about when deciding how groups of words should be written. I could object to the inelegance of the occurrence of two definite articles in so few words, and I could wonder at the ambiguity: who is the thinking person, the author or any reader who finds the book a useful guide?

And arguments can get heated. To deter criticism – rather as Kremlin parades of nuclear missiles averted Moscow’s annihilation – Pinker ends with five things an antagonist should do before engaging. The fifth includes this:

“Psychologists have shown that in any dispute both sides are convinced that they themselves are reasonable and upright and that their opposite numbers are mulish and dishonest. [Footnoting Haidt, J. 2012. The righteous mind: Why good people are divided by politics and religion. New York: Pantheon, and Pinker, S. 2011. The better angels of our nature: Why violence has declined. New York: Viking, chapter 8.]”

Squabbles about being right or wrong are beside the point. The point is style, not grammar. And Pinker concludes by saying that the reasons to strive for good style are:

“to enhance the spread of ideas, to exemplify attention to detail, and to add to the beauty of the world.”

And how to achieve good style? Pinker describes his own writing method, presumably used for this book: rework every sentence, revise a chapter two or three times before showing it to anyone, revise again at least twice in response to feedback, then give the whole book “at least two complete passes of polishing” before it goes to the copy editor for a couple more rounds of tweaking.

Good style doesn’t come easily. And what is the thing that all this revision is trying to get into the writing? Pinker gives plenty of examples of how writing can be improved. In the end, however, all that can really be said by way of defining good style, it seems to me, is that it feels right, and in the case of a revision it feels better than what went before. And this improvement would be accepted by most adults who have English as their first language and who appreciate good writing (the thinking people of his title). Read aloud, to get syntax right so that readers don’t stumble: “laboratory studies have shown that even skilled readers have a little voice running through their heads the whole time.” If it doesn’t sound right it’s not good style.

In one – what is for me significant – respect I find Pinker’s style repellent. I suspect his ears (as he says in relation to a different topic) have “been contaminated by a habit ... to avoid spitballs from the Gotcha! Gang.” It seems to me that his ears have been contaminated in the academic environment by the requirement, appropriate though it may be during a developmental stage of young people’s education, to avoid the sexism of using he to include the female gender, by using he alternately with she, or by using the phrase he or she, or by using their to refer to a singular of either gender.

I agree that sexist language should be avoided, but the right way to do it is to revise until the need for reference to gender disappears. I agree with Antonin Scalia on this (see my review of Scalia and Garner, Making Your Case – The Art of Persuading Judges).

Pinker begins (according to my Kindle app this is p (iv)) by telling us that he will “avoid the awkwardness of strings of he or she ... [by] consistently referring to a generic writer of one sex and a generic writer of another. The male gender ... will represent the writer in this chapter; the roles will alternate in subsequent ones.” But later (p 260) he refers again to science: “Experiments that measure readers’ comprehension times to the thousandth of a second have shown that the singular they causes little or no delay, but generic he slows them down a lot.”

So generic he is bad. Presumably generic she is too. So Pinker’s plan was to alternate bad style chapters with other bad style chapters.

Although he acknowledges that he or she is clumsy (p 256), his plan leads Pinker to do worse by alternating male and female pronouns, even in the one sentence (p 29): “The writer can see something that the reader has not yet noticed, and he orients the reader’s gaze so that she can see it for herself.”

This would have been better: "The writer can see something that the reader has not yet noticed, bringing it to the reader’s attention." And even better: "The writer can bring something new to the reader's attention."

And sillier too is the image one gets of Pinker marking a student’s paper, when he says that (p 28) a “college student who writes a term paper is pretending that he knows more about his subject than the reader and that his goal is to supply the reader with information she needs ...”.

Better would have been: "a college student who writes a term paper is pretending to know more about its subject than the reader and that the goal is to supply the reader with needed information." Even better: "a college student who writes a term paper is pretending to tell the reader something new."

My suggestions are not perfect, and I’m not calling Pinker mulish, but they point to directions for further revisions.

This narrow point of objection to Pinker’s style may not be shared by you (I have done this sort of thing before), and it doesn’t really diminish the value of the book. A good guide for everything except avoidance of sexist language.