Thus, a surgeon performing a dangerous operation would not be reckless, but someone who kills or injures another in a game of Russian Roulette would be, since the action has no social use at all.
In this way, recklessness is worked out by the balance of probability, usually a civil law trait. For example, the above instance of an operation would not be reckless as it is necessary. Indeed, defence by necessity has been employed in such cases, most recently the case of the Siamese twins. .
A final point of difficulty is that of "the reasonable person" versus the defendant himself. In other words, an actus reus may seem obviously reckless to an ordinary person, but would this necessarily be the case for the defendant in question? This is problematic throughout many areas of the criminal law, but the classic example for recklessness specifically is that of Elliott v C. The Divisional Court ruled that the defendant was guilty as the risk only needed to have been obvious to "the reasonably prudent person." Unhelpfully, the court did not define who the reasonably prudent person was, but one is to assume it is the normal person on the street. This is awkward since the defendant may have incapacitating conditions, such as C's educationally sub-normal status in the afore-mentioned case. This precedent has not been adhered to rigorously in other areas of criminal law, especially with regard to provocation in reducing murder to voluntary manslaughter (See, for example, R v Smith.) .
The courts have defined recklessness as falling into two main types - Subjective or Cunningham recklessness and Caldwell recklessness - which further emphasised the "reasonable person" problem.
Subjective recklessness comes from the case of Cunningham in 1957. The defendant planned to steal money from a gas meter in an unoccupied house, but in doing so, fractured a pipe transporting dangerous gas. The gas was diffused and spread to the neighbouring house, where the defendant's mother-in-law inhaled it.
