MARAIS JA:  Few problems have so persistently exercised the minds of lawyers everywhere as liability for omissions in the law of delict (tort). A particularly thorny aspect of the wider problem is the liability of local authorities exercising purely permissive statutory powers of roadmaking and repair when citizens suffer damage as a consequence of the state of a road or pavement. This is yet another case in which these issues arise.
 First, the facts. Respondent, an elderly lady, lives in Mount Nelson Road, Sea Point in Cape Town. She was walking along the pavement of that street towards her home. There were two holes in the tarred pavement which had been there for at least six months. They were approximately fifteen centimetres in diameter and about ten centimetres deep. She stepped into one of the holes, stumbled, and fell. She had been aware of the existence of the holes but “must have been thinking about other things” when she stepped into them. She sustained injuries and suffered loss. Within a few days of the incident appellant (the Municipality of Cape Town) repaired the holes in the pavement. The relevant applicable legislation empowered, but did not oblige, appellant to construct and maintain and repair streets and pavements within its area of jurisdiction.
 Respondent’s claim for damages was upheld in the magistrate’s court. The magistrate made no finding on the question of respondent’s possible contributory negligence, an issue which had been raised by appellant. Appellant appealed to the Cape Provincial Division. A full court (Fagan DJP, Brand et Hlophe JJ) reviewed the applicable law and concluded that the fetters upon the imposition of liability in delict in cases of omission were no longer as rigid as had at one time been supposed. It considered that earlier cases1 decided in this court which accorded a large measure of immunity from action to local authorities which were empowered, but not obliged, to build and maintain streets and pavements, were no longer to be regarded as authoritative in the light of subsequent decisions2 of this court relating to omissions, albeit in admittedly different contexts.
 Writing for the court a quo, Brand J opined that the relative immunity conferred upon local authorities in what have come to be known as “the municipality cases” in this court was inconsistent with the current “legal convictions of the community”3 which require “municipalities to keep streets and pavements in a safe condition”. Having characterised failure to do so as wrongful, the learned judge proceeded to consider whether the failure was attended by fault (culpa) and concluded that it was.
 He acknowledged that, in considering whether or not fault could be attributed to a municipality, account would have to be taken of all factors, including financial constraints, which have a bearing upon the reasonableness or otherwise of the omission. His conclusion was expressed thus:
“It follows from the aforesaid legal principles that appellant’s failure to repair the holes constitutes an unlawful act of omission. The only question is therefore whether appellant was negligent. The uncontested evidence of respondent was that the holes in question had been there for at least six months prior to the accident. The fact that the holes were repaired within two days after the accident, justifies the inference that such repairs did not impose an undue burden on appellant. In the absence of any explanation why the repairs to the pavement were not effected much earlier, I cannot criticise the learned magistrate’s finding that the appellant was negligent. In fact, this was fairly conceded by Mr Binns-Ward in argument.”
 Turning to the question of contributory negligence, the learned judge found respondent to have been equally to blame and reduced the award of damages of R1 500 to R750. The judgment is reported.4 With the leave of the court a quo, given because of the importance to appellant of the principle of law involved in imposing a legal duty to repair streets and pavements upon it, the matter is before this court. An understandable but unfortunate aspect of the case is that there was no appearance for respondent in either the court a quo or in this court. She has abided the judgment of the court and appellant did not seek a costs order against her in either court. It has meant of course that we have not had the benefit of counter argument from respondent.
 The legal literature on the wider topic of liability for omissions generally has burgeoned over the years and has by now reached formidable proportions. Nothing short of a doctoral dissertation can do justice to it all. What follows is a blend of my own observations and what can be gleaned from the more recent cases decided in this and other courts in South Africa and elsewhere, and from the preponderance of legal writing in the text books and journals.
 Society is hesitant to impose liability in law for, as it is sometimes put, “minding one’s own business”. The reticence is reflected in legal and judicial writing by propositions such as no liability in delict for pure (or mere) omissions. The problem with such beguilingly simple propositions is that, however convenient they may be, they are apt, at worst, to mislead the unwary and, at best, to be unhelpful. The proposition that there is no liability in law for minding one’s own business is sound only if, in the eyes of the law, the situation which has arisen, is someone else’s business and not one’s own. But whether that is indeed so is, of course, the very question which has proved so difficult to answer in every age. It is implicit in the second proposition, qualified as it usually is by the use of accompanying epithets such as “pure” or “mere”, that there are omissions which are not of that character. But what kind of omissions those might be, is left unanswered by such formulations.
 Any attempt to decide whether a particular omission will potentially ground liability by merely measuring it against the standard of conduct to be expected of a reasonable person will fail for a number of reasons. First, that test is sequentially inappropriate. It is of course the classic test for the existence of blameworthiness (culpa) in the law of delict. But the existence of culpa only becomes relevant sequentially after the situation has been identified as one in which the law of delict requires action.5 Secondly, the application of the classic test for culpa to the solution of the anterior question is calculated to produce consequences which are likely to be too burdensome for society to acquiesce in shouldering them. The hypothetical reasonable person (diligens paterfamilias) would have to be credited with a reasonable sense of ethical or moral responsibility and a propensity to act in accordance with it. To use his or her likely reaction to the situation as the yardstick by which to measure whether or not action is required by law would be tantamount to converting every reasonably perceived ethical or moral obligation to act into an obligation or duty imposed by law. But that is the very equation against which the law has thus far set its face.
 The instinctive reluctance of society to sanction the imposition of delictual liability on the strength of such an equation is precisely because it is apprehensive about the consequences of simplistically converting moral or ethical obligations into legal duties. It is that fear which provides the impetus for the quest by writers and the courts for a via media between the social inutility of a barren doctrine that denies liability for any omissions and the extravagance of a wholesale conversion of ethical or moral obligations into legal duties. As to the latter, society is simply not prepared to live under so potentially demanding and onerous a legal regime in the area of omissions in the law of delict.
 As to the former, the ways in which the courts sought to escape, Houdini-like, from the confines of the no liability for acts of omission doctrine were many and varied. Initially, the techniques used appeared to many to be casuistic and not linked by any coherent principle. A doctrine of “prior conduct” evolved, the gist of which was that the defendant’s own prior conduct may have been such as to give rise to a legal duty to act. The introduction of “a new source of danger” was an example of such prior conduct. But what of prior conduct that was neutral in terms of creating a risk of harm to others? Or cases in which there was no discernible prior conduct on the part of the defendant? Sometimes a legal duty to act was found to exist because of a particular relationship (“proximity”) which existed between the parties. Sometimes it was found to exist because a duty to act was imposed by statute (despite the fact that the statute did not itself create an independently existing cause of action for damages for its breach). Sometimes it was found to exist because the defendant had control of the property upon which a hazard arose. The list is not exhaustive.
 More recently a much criticised6 doctrine of “general reliance or dependence” has emerged in the Antipodes.7 The thrust of it is that if there is a “general expectation” in the community that a statutory power will be exercised, then even although the statute does not compel the exercise of the power, its non-exercise may potentially ground liability in tort.
 Looking back at the intellectual war of words which has raged for so long in this connection, it is easy enough to discern the battle lines. On one side were those who were averse to what they regarded as timorous incrementalism founded on nothing more than a polyglot and casuistic assemblage of cases thought to resemble one another in one or other respect regarded as significant. Their aversion set them off in search of a readily identifiable and user-friendly principle the application of which would yield predictable and just results. On the other side were those who, while not antagonistic to the search for such a principle, had become convinced that it was hopeless and that their energies should be devoted instead to defending a policy of pragmatic incrementalism reflective of current societal notions of justice. There were also those who hovered in the buffer strip of no man’s land, torn between their philosophic affection for rational principles capable of being consistently applied and their innate sense of caution and appreciation of how frequently unintended and unwelcome consequences flow from well-intentioned attempts to make the law less complex than it is. It is far less easy to decide with which group one should throw in one’s lot.
 Was there a unifying link in the omissions considered in the cases which would provide a coherent and intelligible principle by which to decide whether more than moral or ethical disapproval was called for and whether a legal duty to act should be imposed? It was not always easy to discern one. In the end, this court felt driven to conclude that all that can be said is that moral and ethical obligations metamorphose into legal duties when “the legal convictions of the community demand that the omission ought to be regarded as unlawful”.8 When it should be adjudged that such a demand exists can not be the subject of any general rule; it will depend on the facts of the particular case. It is implicit in the proposition that account must be taken of contemporary community attitudes towards particular societal obligations and duties. History has shown that such attitudes are in a constant state of flux.
 While that attempt to devise a workable general principle by which to determine on which side of the moral/legal divide a duty to act falls has not been universally acclaimed,9 it has been welcomed by most.10 Those who welcome it do so because of its inherent flexibility and its liberation of courts from the conceptual strait jacket of a numerus clausus of specific instances in which a legal duty to act can be recognised. Those who do not are distrustful of the scope it provides for equating too easily with the convictions of the community a particular court’s personal perception of the strength of a particular moral or ethical duty’s claim to be recognised as a legal duty. That is a risk which is not peculiar to this particular problem. There are many areas of the law in which courts have to make policy choices or choices which entail identifying prevailing societal values and applying them. But courts are expected to be able to recognise the difference between a personal and possibly idiosyncratic preference as to what the community’s convictions ought to be and the actually prevailing convictions of the community. Provided that courts conscientiously bear the distinction in mind, little, if any, harm is likely to result.
 The present position regarding omissions in the law of delict is accurately described by Corbett JA (as he then was) in the public lecture entitled Aspects of the Role of Policy in the Evolution of our Common Law and published in  104 SALJ 52. The learned judge of appeal said (at 56):
“Even in 1975 there were probably still two choices open to the court in the Ewels case. The one was to confine liability for an omission to certain stereotypes, possibly adding to them from time to time; the other was to adopt a wider, more open-ended general principle, which, while comprehending existing grounds of liability, would lay the foundation for a more flexible and all-embracing approach to the question whether a person’s omission to act should be held unlawful or not. The court made the latter choice; and, of course, in doing so cast the courts for a general policymaking role in this area of the law.”
 In playing that general policymaking role a court should be mindful of its limitations in diagnosing accurately and prescribing effectively for the ills of society. Some have thought that the legislature is the more appropriate sounding board for proposed extensions of liability in cases when public and private law intersect, as they do in the municipality cases.11 Be that as it may, when a court is required to consider whether a legal duty should be imposed in a given situation the “balance ultimately struck must be harmonious with the public’s notion of what justice demands”.12
 With that prelude I turn to the specific omission in issue in this case. Appellant’s case was argued in the broad rather than with particular reference to the facts of this case. In substance the contention was that the relevant legislation imposed no obligation and cast no duty upon the municipality to build or maintain pavements. It merely empowered it to do so. That distinction, so it was argued, had been regarded as critical in all the municipality cases decided in South Africa. Where such was the case, and in the absence of any antecedent or concomitant act of commission by the municipality which might necessitate a different result, it had been consistently held that no legal duty emanating from the law of delict to repair a street or pavement could arise.
 There can be no doubt that that is indeed the import of the municipality cases in South Africa. So entrenched did the principle become that by 1958 when Moulang’s case13was decided by this court, Schreiner JA felt able to speak of “the general immunity” recognised in those cases and “the high degree of immunity for municipalities in relation to accidents caused by potholes and the like in the surface of streets”.
 Little will be gained by subjecting each of the South African cases decided prior to Moulang’s case to individual analysis. Their import is reflected accurately enough in the judgment in the latter case. Before considering whether their authority has been undermined or terminated by decisions such as those in Regal,14 Quathlamba,15 and Ewels16 it would be as well to be clear as to what it was the municipality cases did and did not decide.
 First, they did not decide that at common law a municipality was absolutely immune from liability and that in no circumstances could it become obliged to repair a road or pavement or fall under a duty to warn of an unrepaired road or pavement.
 Secondly, they did not decide that the relevant empowering legislation per se conferred, either expressly or by necessary implication, absolute or even relative immunity. Nor of course could they have so decided; the legislation was manifestly purely empowering legislation and it was silent on the question of what obligations might arise in the law of delict if damage was suffered as a consequence of a negligent omission to repair a road or pavement ownership of which was vested in a municipality. The fact that there have been and may still be, existing side by side with such purely empowering legislation, other legislation imposing duties of repairs is not sufficient justification for reading into the former class of legislation an intention to exclude such liability as might exist at common law for failure to repair a road or pavement. The priority which would have to be given to the repair of roads and pavements falling within the latter class would be of course a relevant factor in deciding whether or not to impose a legal duty to repair a particular road or pavement falling within the former class or, where a legal duty to repair arising under the common law is found to exist, in deciding whether the omission to repair a road or pavement falling within the former class was culpable. The weight to be assigned to the factor will depend upon the circumstances of the particular case.
 Thirdly, they did not decide that if a municipality chose to exercise its powers of repair, it could not be held liable even if it acted negligently in carrying out the repair. On the contrary, it was recognised that it would indeed be liable.
 Fourthly, they did decide that, absent any antecedent or concomitant act of commission by a municipality which altered the case, the law of delict did not give rise to a legal duty to repair a street or pavement. That conclusion did not rest solely upon the permissive and non-obligatory nature of the relevant legislation and the narrow view taken of the scope of liability in the common law for omissions. It rested at least in part upon policy considerations thought to make it undesirable to impose a legal duty to repair upon municipalities.17
 To what extent, if any, are the cases which have broadened the scope of potential liability in delict for omissions destructive of the municipality cases? They certainly do not expressly profess to overrule them. However, it seems plain that they undermine at least part, and a substantial part at that, of the foundations upon which the “general immunity” doctrine rested in those cases. In so far as the municipality cases proceeded from the premise that “our law of negligence recognises liability for omissions only exceptionally, and more particularly when there has been a previous act of commission on the part of the alleged wrongdoer”,18 they inhibited the courts concerned from enquiring whether, notwithstanding the absence of a legislatively imposed duty to repair or any prior or concomitant act of commission, the legal convictions of the community demanded that a legal duty to repair (or to warn) dehors the legislation should be recognised.
 It is true that in Moulang’s case this court re-asserted the general or relative immunity of municipalities in this area of the law despite declining to investigate, far less decide, what “the better view about liability for omissions in general”19 might be, and that this may suggest that it did not regard the correctness of its narrow view of that liability as critical to the continued existence of that immunity. Nonetheless, once it has been accepted (as it has been) that the premise was indeed erroneous, the authority of the conclusions reached in the municipality cases in regard to any supposed general immunity and the scope of liability for omissions in general must necessarily be considerably diminished. In other respects, the authority of those cases remains unimpaired.
 While the court a quo’s conclusion that it was open to it to re-visit the general or relative immunity of municipalities and, if justification existed to jettison the notion, was therefore correct, I think that, having done so, it was wrong to substitute for it what amounts to a blanket imposition upon municipalities generally of a legal duty to repair roads and pavements. In my view, it has to be recognised that in applying the test of what the legal convictions of the community demand and reaching a particular conclusion, the courts are not laying down principles of law intended to be generally applicable. They are making value judgments ad hoc.
 A minuscule and underfunded local authority with many other and more pressing claims upon its shallow purse, and which has not kept in repair a little used lane in which small potholes have developed which are easily visible to and avoidable by anyone keeping a reasonable look-out, may well be thought to be under no legal duty to repair them or even to warn of their presence. A large and well-funded municipality which has failed to keep in repair a pavement habitually thronged with pedestrians so densely concentrated that it is extremely difficult to see the surface of the pavement, or to take evasive action to avoid potholes of a substantial size and depth, may well be under a legal duty to repair such potholes or to barricade or otherwise warn of them. There can be no principle of law that all municipalities have at all times a legal duty to repair or to warn the public whenever and whatever potholes may occur in whatever pavements or streets may be vested in them.
 It is tempting to construct such a legal duty on the strength of a sense of security engendered by the mere provision of a street or pavement by a municipality but I do not think one can generalise in that regard. It is axiomatic that man-made streets and pavements will not always be in the pristine condition in which they were when first constructed and that it would be well- nigh impossible for even the largest and most well-funded municipalities to keep them all in that state at all times. A reasonable sense of proportion is called for. The public must be taken to realise that and to have a care for its own safety when using the roads and pavements.
 It is not necessary, nor would it be possible, to provide a catalogue of the circumstances in which it would be right to impose a legal duty to repair or to warn upon a municipality. Obvious cases would be those in which difficult to see holes develop in a much used street or pavement which is frequently so crowded that the holes are upon one before one has had sufficient opportunity to see and to negotiate them. Another example, admittedly extreme, would be a crevice caused by an earth tremor and spanning a road entirely. The variety of conceivable situations which could arise is infinite.
 Per contra, it would, I think, be going too far to impose a legal duty upon all municipalities to maintain a billiard table-like surface upon all pavements, free of any subsidences or other irregularities which might cause an unwary pedestrian to stumble and possibly fall. It will be for a plaintiff to place before the court in any given case sufficient evidence to enable it to conclude that a legal duty to repair or to warn should be held to have existed. It will also be for a plaintiff to prove that the failure to repair or to warn was blameworthy (attributable to culpa). It is so that some (but not all) of the factors relevant to the first enquiry will also be relevant to the second enquiry (if it be reached), but that does not mean that they must be excluded from the first enquiry. Having to discharge the onus of proving both the existence of the legal duty and blameworthiness in failing to fulfil it will, I think, go a long way to prevent the opening of the floodgates to claims of this type of which municipalities are so fearful.
 In the present case there is very little in the way of evidence to go on when it comes to deciding whether or not it should be held that the municipality was under a legal duty either to repair these holes or to warn the public of their existence and that its failure to do either was negligent. However, there is just enough to warrant a finding that it was. Sea Point is a densely populated suburb. The pavement abutted on residences and would have been in constant use. There were two holes in close proximity to one another and they were not shallow. There was also a pole near the holes from which a wire cable ran which was attached to the pavement in the vicinity of the holes. It had the effect of shepherding a passer-by in the direction of the holes. The pavement was relatively narrow. The holes had been there for many months. No evidence was given on the municipality’s behalf. In this court Mr Binns-Ward adopted the position that unless the immunity conferred by the municipality cases was re-affirmed, the municipality accepted that it would be liable. In the circumstances, it is unnecessary to subject to any further scrutiny the factual foundation for the existence of a legal duty and a finding that there was culpa in failing to fulfil it.
 The appeal is dismissed. There will be no order as to costs either in this court or in respect of the application for leave to appeal.