1. [1982) AC 341.

2. [1983) 1 WLR 939, 2 All ER 1005, 77 Cr App R 103.

3. (1984) 79 Cr App R 334.

4. Seymour (1983) 2 AC at 507F.

5. That the direction does not apply to crimes of malice was held by a Divisional Court in W (A Minor) v Dolbey (1983) Crim LR 681. The court produced the powerful argument that although it had been said that statutory malice could take the form of recklessness, the authorities had always defined recklessness for this purpose in subjective terms; so it remained the law that malice extended to recklessness only in the sense of subjective recklessness. The dictum in Seymour n 4, was not cited.

6. Law Com No 143 cl 22.

7. [1957) 2 QB 396. The subjective theory of recklessness is of course older. Its first statement in our juristic literature was by Sir John Salmond, Jurisprudence and Professor Kenny, Outlines of Criminal Law both of whose first editions appeared in 1902. Salmond's discussion was the fuller, but it was nevertheless brief; and Salmond underestimated the value of the concept, regarding the distinction of recklessness from inadvertent negligence as ‘of little practical importance’. The subjective definition won support from other writers, including philosophers (see particularly Brady in 43 Mod L Rev 381), and increasingly from the courts. Although there was by no means unanimity, a stream of judicial authority in its favour can be dated at least as far back as 1875; see Williams, The Mental Element in Crime (Jerusalem and Oxford 1965) Ch 2. The lords accepted it in Morgan (1976) AC 182 and also in two civil cases: Derry v Peek (1889) 14 App Cas 337 and Herrington v British Rlys Board (1972) AC at 898F-G (Lord Reid), 921F (Lord Wilberforce), and 928C (Lord Pearson). Geoffrey Lane LJ (as he then was), giving judgment in Stephenson (1979) QB at 73, cited the latter case and said: ‘It would be strange if the meaning of “reckless” in the Criminal Damage Act 1971 were harsher towards an accused person than its meaning in the law of tort is to the defendant’. Lord Diplock did not cite Herrington in his judgment in Caldwell though he had taken part in it.

8. [1986) AC 455.

9. Cp Williams, Criminal Law: The General Part (2nd edn) p 59. An illustration is Chief Constable of Avon and Somerset v Shimmen (1986) 84 Crim App R 7; see particularly the note by J C Smith in (1986) Crim LR 800 on the survival of this rule since Caldwell and on Lord Diplock's confusion of the issue in Lawrence. In some cases it would be reasonable (justifiable) to run an insubstantial risk. For example, suppose the law makes it an offence to deal in uncustomed goods, knowing that they are uncustomed or being reckless as to the fact. The public interest in the freedom and security of transactions would indicate that the offence would not be committed if the defendant knew merely that there was a chance of illegality, particularly if there were no reasonable means open to him of ascertaining the facts.

The American Law Institute's Model Penal Code, POD s 2.02(2)(c), requires that the risk must be ‘substantial and unjustifiable’, and that the disregard of it must involve a ‘gross deviation from the standard of conduct that a law-abiding person would observe’. I disagree with the requirement of substantiality as a matter of general principle, for the reason stated in the text, and doubt the necessity for requiring a gross deviation. Certainly this would be too restrictive for recklessness in some public welfare offences. The proposed new Canadian code is also defective on the point: it defines recklessness in terms of known probability (Law Reform Commission of Canada, Report 30, p 100, s 11(c)).

10. Above, n 9.

11. Textbook of Criminal Law (2nd end) p 656.

12. The unfortunate decision in Masih (1986) Crim LR 395 should not be followed.

13. Elliott v C, n 2 above. In Stephenson (1979) QB 695 the CA had held that a person could not be convicted of arson if by reason of schizophrenia he did not realise the danger of what he was doing. A conviction would have been just as scandalous as that of Miss C was; but it seems perfectly clear on principle that the result in Stephenson should have been an insanity verdict. This was not available in Elliolt v C because the trial was summary; but a conviction should have been out of the question in either case.

14. Westminster City Council v Croyalgrange Ltd (1986) 1 WLR 674, 2 All ER 353, 83 Cr App R 155, following Lord Diplock's dictum in Woolmington (1935) AC 462. Lord Bridge spoke to much the same effect in Croyalgate but he required the rebutting evidence to ‘originate from the accused’, which was surely a mistake. Almost always the defendant would have to give evidence to support a submission of no case, but evidence supporting the defence might occasionally come from a prosecution witness (eg, admitting that the defendant was abroad during the whole of the time when he might have acquired knowledge). Such evidence could ‘suggest absence of knowledge’ within Lord Brightman's condition (3). I proposed a change in the law in (1988) Crim LR 97.

15. Certainly recklessness cannot be transferred from person to property, any more than intention can. Cp the words of Eveleigh J, interpreting the meaning of recklessness in a civil case: ‘If all that can be anticipated is the spilling of a cup of tea over someone's dress, it does seem wrong that the [defendant] should be blamed [on the basis of recklessness] for unexpected personal injuries’ (Goldman v Thai Airways (1983) 1 WLR at 1196H).

16. Clarence Wilson (1983) 1 WLR 356, 1 All ER 993.

17. [1983) 2 AC 131.

18. Satnam (1983) 78 Cr App R 149.

19. ‘Belief and ‘believe’ should be defined in the code.

20. (1889) 14 App Cas 337.

21. Elliott v C, n 2 above.

22. CI 26(2).

23. When discussing the specific case of intoxication (1982) AC at 355F), Lord Diplock formulated the rule in the conditionally subjective terms now adopted by the Draft Code (which had appeared in the Model Penal Code of the American Law Institute) - ‘a risk of which he would have been aware had he been sober’ (emphasis supplied). Regrettably, in his model direction relating to recklessness in general, Lord Diplock stated the general rule in objective terms - ‘when he does the act he… has not given any thought to the possibility of there being any such [ie, obvious] risk’.

24. An illustration of anger against an object counting as recklessness is Parker (1977) 1 WLR 600, 2 All ER 77. The CA there tried to reformulate the subjective definition of recklessness to uphold a conviction, but unfortunately did not confine itself to the specific situation of anger, which was the one it had to consider. On the whole question see R A Duff in (1982) CLJ 273, and my reply, ibid 286. Mr Duff effectively criticised various suggestions I made for producing a satisfactory definition of recklessness, but I remained unconvinced that his own suggestions were workable. I see the force of a distinction he draws (at p 280) between (1) the person who angrily assaults another in dangerous circumstances, and by reason of his anger does not realise the danger he is causing to the victim of the attack, and (2) the driver who angrily gets out of his car in order to remonstrate with another driver, and by reason of his anger does not consider the danger he is causing to a passing cyclist. ‘While [in (1)] the risk to his victim's life is an integral aspect of the assailant's intended attack, [in (2)] the risk to the cyclist is only contingently and coincidentally connected to the motorist's intended action’. Accepting this, I have met the point by including in my formulation (see text above) the words ‘anger directed against the person or thing that he harms’.

25. Textbook of Criminal Law (2nd edn) p 475. In the case referred to there was an element of intoxication as well; but even under the intoxication rule as formulated in the Draft Code the defendant could argue that it was the excitement and not the intoxicant that blinded him to the risk, ie, that in the excitement of the moment he would not have realised the risk even if he had not taken an intoxicant. From this point of view the facts are borderline, and whether the jury convict or not would depend on their attitude. A conviction would not be against principle, though it might look hard.

26. This solution was proposed by Gerald H Gordon, after a spirited attack upon the utility of the notion of recklessness, in 17 Crim LQ 355 (Can); cp Griew in (1977) Crim LR 100–101.

27. [1967) 2 QB 981.

28. I formerly accepted these reasons: see (1981) CLJ 278-282 and 132 NLJ 313, 336. Cp Syrota in (1981) Crim LR 658; (1981) CLJ 268-272; 132 NLJ 290, 314. I now have much less confidence that a court will accept them.

29. Above, n 2.

30. Elliott v C and R (Stephen Malcolm), nn 2 and 3 above.

31. Chief Constable of Avon and Somerset v Shimmen, n 9 above.

32. See the argument stated at length by Syrota in (1982) Crim LR 97, and the criticism by R A Duff in (1982) CLJ 273.

33. See BJ Mitchell in 150 JPN 390; Richard Taylor in 137 NLJ 232. Mitchell points out that the rule in Caldwell is now applied to manslaughter, and suggests that this means that a person who has negligently ruled out a real risk, and who kills in consequence, is no longer guilty of manslaughter. But the possibility of such a conclusion would be likely to frighten a court off accepting the ‘ruling out a risk’ principle.

34. A Divisional Court in Shimmen's case, n 9 above, expressed itself as ready to accept a defence of ruling out risk, but would not accept the defence in a case where a reasonable man would still have seeen a risk, which seems to bring the question back to negligence. The judgment is unsatisfactory for this reason and also because it fails to decide whether the defendant thought he had ‘eliminated’ the risk or merely ‘minimised’ it.

35. In Herrington v British Rlys Board (1972) AC at 934M Lord Diplock said: ‘This House has since 1966 abandoned its former practice of adhering rigidly to the ratio decidendi of its previous decisions’.