Emptio-Venditio (Sale)

Sale is the commonest contract and, in that sense at least, the most important. I propose to deal with it quite fully, at the cost of cutting down the space available for the others.

If you were a buyer and were driven to sue your seller, this is the formula which you would use:

Quod Aulus Agerius de Numerio Negidio hominem quo de agitur emit, qua de re agitur,

Whereas Aulus Agerius bought the man who is the subject of the action from Numerius Negidius, which matter is the subject of the action, quidquid ob eam rem Numerium Negidium Aulo Agerio darefacere oportet ex fide bona,

Whatever on account of that matter Numerius Negidius ought to give to or do for Aulus Agerius in good faith,

eius iudex Numerium Negidium Aulo Agerio condemnato; si no paret absolvito. for the value of that let the judge condemn Numerius Negidius to Aulus Agerius; if it does not appear, let him absolve.^[16]

The pattern of this, the formula of the actio ex empto, was given in the edict. All you had to do was to accommodate it to your own facts, putting your own and your seller's name instead of the characters N.N. and A.A. (�Numerius Negidius' and �Aulus Agerius': names made up from negare—�to deny or defend'—and agere—�to plead or claim') and your own res for the slave here supposed to have been bought. If you were seller yourformula, now of the actio ex vendito, would be exactly the same with the parties turned about and �sold' used for �bought'.

You can see that this programme reflects a claim, in direct speech, on these lines: �I bought the slave from you and you ought in good faith to...' Before the iudex the party will have to establish the fact of the purchase and then show that, according to the implications of good faith as settled by juristic interpretation, something or other is indeed owing.

i. The demonstratio

�Whereas X bought the res from Y.' This is a pruned down version of the clause which in the language which describes the syntax offormulae is called the demonstratio. What facts have to be the case before it can be said that one person has bought something from another? What would the plaintiff have to prove? This is the same as asking, What is the legal definition of emptio-venditio?

The short answer is that there must have been an agreement to exchange a res for a price. There are four elements in that statement. I hope it is not merely perversity which makes me take them in reverse order: (i) price, (ii) res, (iii) to exchange, and (iv) agreement.

A. Price (pretium) Before the judge you show that you agreed to give your horse for the slave. Does that show that you bought him? In other words, is barter within the definition of sale?

This was a point disputed between the schools. The Proculians maintained that barter and sale had to be different contracts, because in sale different obligations lay upon the parties. Only by insisting on a money price could the law distinguish the buyer from the seller for the purpose of distributing these different incidents. The difficulties start with the action itself. �Whereas he bought'—who should say that? This seems an irresistible argument. Nevertheless the Sabinians maintained the contrary. And Caelius Sabinus seems to have thought that the necessary discrimination between buyer and seller would be achieved by asking which had offered something for sale and which had bid.

Gaius states this dispute in the present tense, as though the point was still unsettled. And he loyally gives prominence to the Sabinian posi­tion. But the rule he propounds at G.3.141 is actually that of the other school: pretium in numerata pecunia consistere debet (the price must consist in money).

Even those who take this position accept that sale finds its origin in barter. In D.18.1.1 (Paul, 33 On the Edict) Paul speaks of the ineffi­ciency of barter as being the stimulus for the invention of money as a medium of exchange. His account seems to assume that coined and minted money is the first step, whereas the Roman evidence indicates an intermediate stage in which uncoined bronze served as the universal medium of exchange. Whence the survival of scales in formalities �per aes et libram.

The decision to insist on a money price was not pushed further than necessary to achieve the discrimination for which the Proculians argued. What we call â€?part-exchange’ (my car plus £1,000 for a new vehicle) remained within sale. There is a hint of doubt about the transaction which is really an exchange into which one party injects a few pounds as a makeweight. Perhaps the safest thing is to say that the price had to be in money but need not be wholly in money, so long as the non-money element did not bulk so large as to destroy the money's power to distinguish buyer from seller.

Final commitment to a price naturally comes last in the negoti­ations for a sale. For other matters, for example the question as to who must answer for defects or repairs, will be reflected in the price. The more responsibility accepted by the seller, the higher the price. So the fixing of the price seals the entire deal. The Romans took that as the moment at which the contractual bond was formed. G.3.139:

Emptio-venditio contrahitur cum de pretio convenerit, quamvis nondum pretium numeratum sit, ac ne arra quidem data fuerit.

Sale is contracted when there has been agreement on the price even though the price has not been paid over and even if no arra has been given. For what is given as arra is evidence of sale contracted.

Then D.18.1.2.1 (Ulpian, 1 On Sabinus):

Sine pretio nulla venditio est: non autem pretii numeratio, sed conventio perficit sine scriptis habitam emptionem.

There is no sale without a price: and it is not the payment of the price but the agreement on the price which perfects a sale (made without writing).

Mention of arra in the passage from Gaius and the interpolation of sine scriptis habitam in the bit from Ulpian remind us of the discussion we have already had, about Justinian’s innovations.^[19] No need to go into that again. It is all very well to say that agreement on the price concludes the contract. But when is agreement on the price itself complete? The answer is, when the price is certum. G.3.140: Pretium autem certum esse debet (But the price must be certain).

What amounts to sufficient certainty? The requirement of certainty does not mean that the price has to be ascertained in figures, only that a formula must have been agreed which will yield the figures without further negotiation. There is a maxim which expresses this, though it makes circular nonsense if you look at it too carefully: certum est quod certum reddi potest.^[20]

If I buy the sweets in a jar at 1 penny per sweet or 20 pence per quarter, weighing and counting will have to be done to reveal the price in figures. But that is all right so far as making the contract is concerned. An aunt gives you Walker’s Principles of Scottish Private Law for Christ­mas, just after you had bought it for yourself. I might say �I’ll give you as much as she paid for it’ or �... as much as you paid for yours’ or �... half as much as she paid for it’ and so on. These too are formulae impressed on the price which fix it sufficiently according to the Roman rules.

Suppose the agreement was that you would sell at a price to be fixed by a third party: quanti ille aestimaverit (for the sum set by X’s valuation). This caused trouble to the classics and Justinian settled the argument by one of his Fifty Decisions, C.4.38.15. There is a shorter version at J.3.23.1. He makes the sale conditional. If X names the price the contract becomes operative; if he fails to do it then the sale is a nullity as though no price had ever been agreed. A decisio supposes a contrary view. What was it? The version in the Codex reveals that some were prepared to argue that if the named X failed to act, the formula pressed on the price might be construed as �a reasonable price', as though not X had been intended (one given person) but simply a representation of reason, X as a reasonable man. C.4.3.15 says, towards the end, that what had to be got rid of was �guesswork as to whether the parties agreed the term with an eye to one certain person or to boni viri arbitrium, a sound man's judgement'.

The view overridden in this legislation has interesting implications for agreements to sell �at a reasonable price', �at a fair price'. The difficulty with the named third party was whether he should be construed to have been named as a representative of reason.

Classical law knew nothing of control of prices by any doctrine which required the price to be fair, the exchange of value roughly equal. The price had to be verum, genuinely intended to be demanded as opposed to merely colourable, pretended. A transaction dressed up as a sale but actually a gift would be void. But a genuine sale for a very cheap price or a very expensive price would be upheld. Only in very late law, in an attempt to curb exploitation of the weak by greedy potentes (magnates), did there emerge a doctrine of iustum pretium based on what was called laesio enormis. The object was to protect the weak who were being bought out of their land. They could set aside the contract if the price was less than 50 per cent of the market rate.

B. Res (subject-matter, or thing) At D.18.1.8 pr.^[21] Pomponius says �Nec emptio nec venditio sine re quae veneat potest intellegi (You can never make out a purchase or a sale if there is no thing being sold).' This principle must be approached with caution. Sale is always of something. It must have a subject-matter. But if you read the requirement of a res quae veneat too literally and concretely you can go wrong. A carpenter can sell �the next chair I make', and the contract is good even though the res is as yet only a twinkle in his eye. And it is possible, if an exercise of construction shows that the parties so intended, for the subject­matter to be no more than a chance or speculation. �All the gold in your land.' None is found. You should not conclude that the sale is necessarily void.

Just as the price, so the res had to be certain. If you complained that the defendant had sold you �part of his field' or �some of his wine' you would merely reveal that your negotiations still had some way to go. But what counted as sufficiently certain? Alternatives were all right. If you showed you had bought �Daisy or Buttercup' the sale was valid and, if nothing was said to the contrary, the seller could pick which to give. And selections from a given stock were all right: �Twenty litres of olive oil from your vat' or �A thousand bricks from the clay on your land'. But what about descriptions which failed to state the corpus from which the res was to come? Suppose we agree a price for 10,000 kilos of Egyptian grain, or just of top quality grain. Is this sale? The dominant view is that it is not. If your formula said �Whereas A. A. bought from N.N. such and such a quantity of grain', Numerius Negidius would be able to object that the sale was generic and as such no sale at all in the law. No emptio generis. The argument is chiefly e silentio. Respectful scepticism is perhaps the best position. If I take cloth to a tailor to have made up into a suit, that is hire; but if I order the suit and do not myself provide the cloth, that is sale. But it is not said that he must have the cloth at the time of the order. So the contract for the suit is, in its own way, by generic description. And where is the line between �100 litres ofBordeaux Superieur', �100 bottles of Chateau Cissac 1975' and �a litre from your vat'?

Suppose the res on which the parties focused their attention never comes into existence: �The yield of this olive tree's next harvest', �This girl's baby when it's born', �the next catch of fish'. There is no fixed answer. It depends on the right construction of the deal. Did they mean �the yield if there is one' or �the hope of a yield'? If the former, emptio rei speratae, the sale is conditional on there being some res; if the latter, emptio spei, the spes (the hope, or gamble) is the res, so that the contract binds even if no yield ever happens. If the seller interferes with the risk, as for example if he will not till the soil, he becomes liable for that. Hence, if you buy next year's crop and I refuse to sow or tend it, I become liable ex empto notwithstanding there being no res.

What if the res never existed? You sell me your sunken treasure ship out in the bay. There is no such ship. You have confused it with another case, without dishonesty. The easy answer is to say �no res, no sale', but the solutions applied to emptio spei and emptio rei speratae suggest an exercise of construction. Did I buy an adventure, including in the gamble even the existence of the ship? If yes, I must pay, and can claim nothing for my expenses. If we assumed the existence of the ship, outside the gamble, either its existence was a neutral condition or you undertook responsibility for its existing. If the latter I must be able to sue you, for interfering with the risk. I will say �whereas I bought the ship... ' even though the ship never was there at all. This is a bit of a guess. The texts are not clear.

What if the res existed but perished before the contract was con­ducted? The sale appears to be void. D.18.1.15 pr. (Paul, 5 On Sabinus):

Et si consensum fuerit in corpus, id tamen in rerum natura ante venditionem esse desierit nulla emptio est.

Even if there is agreement on the physical identity of the thing, the purchase is void if it ceases to exist before the sale.

There are two kinds of difficulty about this proposition. What consti­tutes �ceasing to exist'. Some cases are easy. A book which has gone up in flames no longer exists. Homes leave the land behind. Fruit which goes bad is still there though suitable for the compost heap not the table. The Roman texts do not come up with any very satisfactory test. We will recur to this under �risk'. Secondly, what is the status of the proposition? Is it an inflexible dogma? Or just the commonest construction, in the absence of contrary intents? If it were a dogma it would be impossible within the contract of sale for the seller to warrant the continued existence of the res. So probably nullity simply reflects the normal case. If you sell me the horse at home in your field, the usual construction will be that our deal is conditional on the existence of the horse. Only in exceptional circumstances will I gamble on its existence (i.e. agree to pay whether it is alive or dead) or you warrant its existence (agree to pay my losses arising from your non-delivery).

Suppose the slave you bought turns out to be free. Or the land turns out to be dedicated to the Gods, as for example a burial place. These things exist but they are removed from commerce and ownership. So in a sense they do not exist, not for business purposes. Can you substantiate the demonstratio of the actio ex empto so as to get a remedy that way, or will the defendant be able to answer that the sale is void? The original position seems to have been that the rule was void if the res was extra commercium. But that reduced the range of remedies. At least in the case of the free man the classics therefore favoured making the sale valid and the seller liable. The position of other res extra commercium was later assimilated to that of the free man, probably only by Justinian. The assimilation is not perfect, though the Institutes assert that it is intended to be (J.3.23.5):

Loca sacra vel religiosa, item publica, veluti forum, basilicam, frustra quis sciens emit quas tamen si pro privatis vel profanis, deceptus a venditore, emerit, habebit actionem ex empto, quod non habere ei liceat ut consequatur quod sua interest deceptum non esse. Idem iuris est si hominem liberum pro servo emerit.

Knowingly to buy sacred and religious places, or public ones such as a forum or basilica, is merely to beat the air. But if you are misled by the seller and buy them as private or secular you will have an action ex empto. The ground of the action will be, that you are not being allowed to possess, and the measure of recovery is your interest in not having been misled. The same law applies where you buy a free man as a slave.

What this does is to treat sales of res extra commercium on the model not of res extinctae but res alienae. That is, the analogy of non-existent things is given up, and the analogy of things belonging to third parties is followed instead. We have not looked at res alienae.

Suppose you sell me a cow belonging to your neighbour. Even if you mancipate her to me she will still belong to your neighbour, though if all the circumstances are right maybe I will usucapt her after a year has passed. During the year he reclaims her, as is his right. I must come against you. Can you say the sale was void? No. D. 18.1.28 (Ulpian, 41 On Sabinus):

Rem alienam distrahere quem posse nulla dubitatio est: nam emptio est et venditio: sed res emptori auferri potest.

There is no doubt at all that anyone can sell something belonging to another. Indeed there is then purchase and sale: but the res can be taken away from the buyer.

So the buyer cannot keep the res but does have his contract to rely upon against the seller. We will recur to the nature of the contractual remedy when we consider the obligations generated by the contract described in the demonstration

By way of summary, note that we have considered a series of, as it were, special cases or problem cases in relation to the res vendita. These labels will serve to call them to mind: (a) emptio generis, (b) emptio spei and rei speratae, (c) emptio rei extinctae, (d) emptio rei extra commercium, (e) emptio rei alienae.

C. Exchange We said that sale is an agreement to exchange a res for a price. At this point the focus falls on �to exchange'. In 99 cases out of a hundred what happens is that the seller gives up ownership of the res and gets the price instead. Is it completely safe therefore to substitute for �to exchange a res' the words �to give ownership of a res'? Technic­ally it is not quite right.

First, the Roman seller is not understood as undertaking to make the buyer owner. It is not part of the definition of sale that ownership must pass and it is not part of the seller's package of obligations that he must make it pass. After all we have seen that sale of a res belonging to a third party is valid.

The seller's part is to convey the thing. That is, to do what is necessary to transfer such ownership if any as he has, by traditio or mancipatio. These conveyances are subject to the maxim nemo dat quod non habet. They cannot constitute new titles in alienees. If the alienor is owner, they serve to carry that ownership across. If not, nothing happens. Conveyance is not necessarily successful. Bearing that in mind, we can say that for �to exchange a res for a price' we should say �to convey a res in exchange for a price'.

Secondly, it still counts as sale if the buyer is already owner and the transaction is intended only to obtain possession or to obtain release from a condition under which his ownership is threatened (D.18.1.34.4;^[22] D.18.1.6.1).^[23] These holdings are made against the back­ground of the rule that purchase of one's own property is a nullity (D.18.1.6pr.)^[24] These rare cases make it unsafe to say even that the seller must do what he can to divest himself of ownership.

The negative proposition is certainly true, that the transaction can­not be sale if the seller is to retain ownership, though he may retain a lesser interest such as a usufruct. There is nothing wrong with a sale of a farm or flock deducto usufructu, with a usufruct reserved for, say, three years.

Perhaps it is safe to say that the seller's part is so to convey the res to the buyer so that he retains no interest in it, save such interest less than ownership as is reserved to him. This is what one is driven to by the need to avoid unintended exclusions and distortions. The common sense version is always too narrow. The artificial version is repugnant in other, less dangerous ways.

If I give £20 for one week's use of your car, that is hire, not sale. Similarly if I hire your services as a mechanic. We do not say that I buy the slice of time in your car or your own labour-over-time. If we wanted to compress sale and hire into one we could take that line. But we do not. And nor did the Romans. The two are kept apart by declining to see enjoyment-over-time or labour-over-time as an inde­pendent res. All the same the line between the sale and hire could be very tricky. The difficult cases are most conveniently kept till we come to locatio-conductio itself.

D. Agreement Before the judge your opponent may say that, though you have described a transaction within the notion of sale, he never agreed to it. If that is true he never sold and you never bought. We need not deal with the simplest versions of this. For example, he may maintain that he never dealt with you at all or that he flatly rejected the terms which you say he accepted. The subtler versions of the same thing are those where he says he was mistaken or that the agreement was only conditional. I shall take conditions first.

(a) Conditions. Suppose we agree to the sale of a horse at such and such a price �if he wins the race next Tuesday'. Before Tuesday I tell you that I have changed my mind. And then he wins, and you sue me. There are two views of the matter: no contract until the outcome is known, or contract binding but suspended till the outcome is known. On both views if the condition fails there is no contract: once the horse has lost, the deal is off. The difference is as to the situation in the run-up to the fulfilment of the condition. If there is no contract the parties can get out; if there is a suspended contract they must abide the event.

There are traces of the first view but it did not prevail. In the classical law a conditional agreement was binding. The presence of the condi­tion had an effect on the rules about risk but did not allow the parties to change their minds. They were bound, at least in the sense that only failure of the condition would release them. What if a party took steps to make sure that the condition failed? For example, I refuse to allow the horse to run in the race. There is a doctrine, usually labelled pro impleta, whereby a condition can be regarded as having been fulfilled if frustrated by one party. But this only works when the condition can be waived by the plaintiff without damaging his own interest. �I will buy if you redeem the mortgage.' Clearly I am never going to waive that condition, should you refuse the redemption. If the construction of the transaction is such that you are taken to have promised to race the horse or redeem the mortgage I can have an action for breach of that term.

Not all conditions are suspensive. In that sale of the horse we might have used words, the proper construction of which would be that the sale was to come into immediate effect subject to being resolved if the horse lost the race. The business of unpicking the legal consequences in that event is complex, and the details are disputed. But it is more obvious in the case of a resolutive condition that there is no question of withdrawal pendente condicione: the contract is pure, subject to a question for the future.

(b) Mistake. This is always a difficult topic, chiefly because the inclin­ation to offer relief to the mistaken encounters a fear that people generally will abuse such relief, concocting mistakes ex post and upset­ting all kinds of transactions as soon as they begin to feel disappointed. And, distinct from this, there is another fear peculiar to commercial life, namely that the market will not work if those with bad judgement are not driven to the wall. The premise being, that the avoidance of mistakes is precisely the quality which marks the successful and efficient businessmen. So there are pressures both ways, for and against relief, and that is what makes for puzzles and artificialities in the law.

Sale requires agreement (consensus) in order to come into being. It follows that a party who can show that there was still disagreement between him and the other (dissensus) shows that no contract was made. If he is saying â€?You know perfectly well that I never agreed to sell for £20,’ he does not seem to be raising an issue of mistake. But if he is saying â€?Though I see you thought differently, in fact I never did agree to sell for £20,’ then words such as â€?mistake’, â€?misunderstanding’, â€?error’ and so on do appear to be appropriate. But dissensus is more exact. The word does not matter, so long as the idea is right. Which is, â€?We differed,’ â€? We had different views of the matter.’ And this must be contrasted with vitiated consensus where one party says â€?I thought so too.’ It is the reaction of the second party which distinguishes dissensus from flawed consensus. â€?Well, I didn’t!’ makes it dissensus. â€?That was my view too!’ makes it consensus on the basis of a shared mistake.

The Roman approach made the contract void if there was dissensus as to res (I meant Daisy, you were thinking of Buttercup), price (I meant £100, you thought £50) and negotium, the nature of the deal (I meant to sell, you intended to hire). But difficulties were felt when there was dissensus as to qualities of the res (I thought it was fire­resistant, you did not). It is here that relief for mistake can threaten bargains and also here that mistake can present itself as flawed consensus as well as dissensus. The opinions of the jurists seem to have varied from a determination to give no relief at all to a willingness to avoid the contract for a very fundamental mistake. Ulpian appears to have favoured relief for fundamental mistake of quality, subject however to the spongy notion of fundamentality being crisped up by a mech­anical test, itself not very satisfactory: was there error in substantia, mistake as to the materials? The leading text is D. 18.1.9-15, part of which is D. 18.1.9.2 (Ulpian, 28 On Sabinus) which reads as follows:

If there is no error as to the thing itself (si in ipso corpore non erratur) but rather a mistake as to substantia, is there emptio-venditio or not? As where vinegar is sold for wine, bronze for gold or lead, or something else similar to silver, for silver. Marcellus in book 6 of his Digesta says the sale holds good, because there is consensus as to the thing itself though error as to the materia. I myself agree as to the wine since the essence is the same, at least if the vinegar was indeed soured wine. Yet if it was not sound wine but vinegar ab initio, one thing would seem to have been sold for another. For the rest I think the sale is void as often as there is a mistake as to material.

It seems that he took the view that the same test should apply to cases of flawed consensus as to dissensus (D.18.1.14).¹⁰ That seems sensible. There is only one reason for handling flawed consensus differently, namely that relief is somewhat less contrary to the nature of bargaining when both share the same mistake. On the other hand it is difficult to make much of that since it lies too easily with the other party to choose whether to say he shared the mistake or knew the truth all along.

When is it in a party's interest to try to avoid the contract for mistake? Usually it is a ploy used by a buyer to defeat an action for payment and non-acceptance. After he has found out the truth. Some­times a seller will want the contract void so as to keep silver he thought was lead; or to resist a possible claim for damages if silver turned out to be lead. This all happens within the actions on the sale. There are other contexts. There may be a question of recovering a price already paid by condictio or of arguing in a vindicatio that property in the res never passed.

ii. The intentio

�Whatever in that account Y ought to give to or do for X in good faith...' This is the intentio of the formula, the part which expresses the plaintiff's principal contention. Up to now we have been talking about the demonstratio. The question has been, What facts serve to substantiate the assertion that a sale has happened? Now we are concerned with the consequences of that fact. For the moment we will retain the view­point of the buyer's action rather than the seller's. What is it that a seller ought on account of a sale to give or do in good faith? Or, What is the package of obligations on which a buyer can rely?

A. Before delivery The seller's obligations are to take care of the res till delivery and to deliver it. Suppose I buy a horse from you. Knowing that you are going to give it up to me in a day or two, you mistreat it. You make it gallop for miles and miles. Until it drops. Of course, you did not mean that to happen. You just stopped taking care. And now you will be liable for non-delivery, just as though you had wantonly changed your mind.

10

Ulpian, 28 On Sabinus.

It is always unhelpful to talk of an obligation to take care unless mention is made of a standard. Does he have to be very careful (to take all the precautions that can be taken) or moderately careful (to take such precautions as are taken by the generality of mankind)? And so on. There are many ways in which standards can be expressed. The common law continually invokes �the reasonable man' and, slightly less prominently �the prudent businessman'. One standard used by the Romans was that of the bonus paterfamilias. Literally he is a �good head of the family', but �prudent property-owner' probably catches the idea better. And in fact the modern reasonable man is taken to behave much in the same way as the notional paterfamilias who set this Roman standard.

The Roman seller certainly had to show before delivery of the res the care which would have been shown by a bonus paterfamilias. He would be liable if damage or destruction happened for want of that standard of care. Just possibly the classical law expected even more of him. He may have had to answer for custodia (safe-keeping). If so he would have been liable, whether actually careless or not, if damage or destruction happened, unless they happened because of vis maior (irre­sistible force) such as flood, earthquake, robbery. For example, one who had to answer for custodia remained liable if the res was lost by theft (as opposed to robbery, which is theft with violence) even if he had taken such precaution against thieves as a prudent proprietor would have taken.

Suppose that a disaster occurred despite the seller's having discharged his obligations of care. The horse was struck by lightning. Can the buyer sue for failure to deliver? He cannot. To say otherwise would make the obligation to deliver absolute and the obligation of care strict. For events of this kind the buyer has no more cause for complaint than if he caught measles. It is just bad luck. He will have to pay and get nothing. Just as though his own horse had been struck rather than a horse destined by contract to become his. But there is a qualification. The disaster must have happened after the point at which the risk (periculum) passes to the buyer. In most cases the moment at which the risk passes is the same as the moment at which the contract becomes binding. But not in all. An excursus on the risk of accidental damage would unbalance the discussion at this point. I shall deal with it as a separate topic later.

Suppose that the disaster occurs outside the seller's area of respon­sibility but is also not an act of God. As, for instance, if a third party robs the seller of the horse. Here the buyer has no claim against the seller for his failure to get the horse, but he will want to sue the third party. The seller is therefore put under an obligation to assign the actions against the third party. He has, so to say, to deliver the actions instead of the horse. And he will be liable ex empto if he does not.

One way of summarising this is to say that the seller is liable if he fails to deliver the res and/or actions arising in relation to its damage or destruction; and, on the other side of the penny, is not liable despite failure to deliver either res or actions if disaster overtook it even though he discharged his obligation to look after it between contract and delivery.

B. After delivery There are two things that a buyer chiefly fears. A bad title. And latent defects. Suppose the horse I bought from you turns out to belong to a third party. Or suppose it has a sickness which weakens it and stops it being able to work.

Before dealing specifically with these it is convenient to distinguish two different bases of complaint. There is the fault (or mala fides) basis and the warranty basis. The fault basis of complaint rests directly on the words of the formula which suppose that the seller must behave with good faith. The judge is actually asked what the seller ought to do ex fide bona. And this means that there is a general liability for mala fides as settled and explained by the jurists. Further, this liability is non­excludable. Because of being built into the action. Broadly speaking, the basis of complaint means establishing shabby, unconscientious behaviour on the part of seller, usually by showing that he knew that the buyer was in for trouble.

By contrast the warranty basis of complaint is not dependent on proof of fault. It simply saddles a seller with responsibility for a given state of affairs without imputing any disreputable conduct to him. For example, if a modern seller promises that his carpets will survive under normal condition of fair wear and tear for five years, the reason why he is made liable when they have to be replaced after two is not that he dealt shabbily (though he may have done) but that he promised to answer for their durability. Express promises are the easy case of no-fault liability, but the natura contractus may also be understood to impose the same kind of liability on sellers who do not expressly undertake it.

In one respect the contrast between these two bases of liability is false. For both are referable to the bona fides on which the action is based. That is, where a warranty liability is recognised the reason is that in those circumstances the defendant �ought in good faith' to accept responsibility irrespective of fault. One way of putting this is to say that that �ought in good faith' always requires the defendant to make good the effects of actual bad faith but sometimes requires him to honour liabilities whose origin does not lie in unconscientious behaviour.

(a) Bad title

The seller who knows he is selling a res aliena is liable for his mala fides. For the rest a warranty liability did develop, but not in the form of a liability specifically for failing to make the buyer owner. It was not a warranty of title but a warranty against eviction. Suppose you dis­covered that the horse which I sold and mancipated to you was not mine and hence not yours. If I knew, you could sue me at once, for my bad faith. If I did not know, you could not make me liable unless you were evicted from possession by the third party.

The history of the warranty against eviction seems to go like this. The conveyance of mancipatio carried a built-in warranty. At least from the Twelve Tables the alienee by mancipatio could bring a special action (actio auctoritatis) for double the price if he was evicted. Then, during the Republic, the practice developed of imitating the regime for mancipatio by taking stipulations against the danger of eviction. These would only be necessary where mancipatio was not used. And in the event of eviction the action brought would be on the stipulation, not on the sale. Then, the early classics drew on the general liability for bad faith in this way: they affirmed that good faith required the giving of these stipulations, so that a seller who did not enter into the conven­tional undertakings was by that fact itself guilty of a breach. Finally, but before the end of the classical period, the conventional undertakings were regarded as having been given unless there was an agreement between the parties to the contrary. If there was a pactum de non praestanda evictione (a pact against answering for eviction) the buyer was remitted to the fault-basis of complaint. Without such a pact the seller was liable if eviction happened.

(b) Latent defects

By Justinian’s time the picture is of a general warranty against serious latent defects. But the development is a good deal more complex than in the case of eviction.

There is a preliminary problem about the meaning of the phrase �latent defects’. �Latent’ is the easier part. It means �hidden’ or �not discoverable by such inspection as a buyer could reasonably be expected to make’. But �defect’ is not so easy. Corn which is contam­inated may still be fit for animals to eat. A spear which is brittle may serve very well as a clothes prop. Cloth which is useless for making clothes may be all right as rags or stuffing for furniture. And so on. These are extreme cases. The wine-snob may class as undrinkable bottles which others drain with pleasure. The Rolls Royce mentality regards the ordinary version of everything as vicious. So what is a defect? It is not safe to say that it is a bad quality which, if known, would have reduced the price. Too many sales could be upset if it were a ground for complaint that in use the cooking-pan or pot showed less well than on the seller’s shelf. You are driven towards the notion of merchantability. Is the bad quality such that, declared openly, it would prevent the thing’s being bought and sold under the description and categorisation under which it was sold? Does the defect take the thing out of the commercial category in which it was sold? There is no general test in the texts.

The problem addressed in the last paragraph concerns the case in which no special qualities are ascribed to the res by the seller. If by contrast he says that the slave is literate or skilful as a painter, or that the timber is suitable for boat-building, or that the pots are fire-resistant, then there can be no argument but that the thing is defective if it lacks the quality ascribed to it.

(i) Warranty liability. You will remember that this term is being used to denote liability independent of bad faith or fault of any kind. What was the position in classical law under the actio ex empto? That question implies two exclusions for the time being: the aediles’ jurisdiction and the law of Justinian. Suppose two concrete cases: corn which, contam­inated by poison, kills all the buyer’s chickens, and a chair which, badly glued up, becomes wobbly after a few days’ use.

The first question is, Did the seller undertake to be answerable for the quality in question? Did he give a warranty that the corn was suitable for chickens or that the chair was well made? If he agreed to be liable he will be liable. D.19.1.11.1 (Ulpian, 32 On the Edict):

The first thing to have in mind is that it is the liabilities agreed by the parties (quod praestari convenit) which above all are passed upon in this trial [sc. under the actio ex empto]. For it is a trial based on good faith. And nothing agrees better with good faith than that there should be fulfilled whatever was intended between the contracting parties (quam id praestari, quod inter contra­hentes actum est). But in the absence of agreement the liabilities will be those which the nature of the transaction puts within the scope of this trial (quae naturaliter insunt huius iudicii potestate).

So the position of the seller who accepted liability for the quality is clear. The only difficulty will be in deciding whether he did accept it. There is a line between affirming the existence of a quality and accepting liability for it. But it seems that the practical difficulty of deciding which the seller had done (if he had done either) became less by the end of the classical period. For the high classical jurists seem to have been prepared to treat a seriously intended representation, as opposed to mere puffery, as a warranty. They reached this position by developing the notion of fault-based liability to the point at which fault was inferred from the making of the representation, that is, by so denaturing the requirement of fault that the liability became warranty­based. D.19.1.13.3 (Ulpian, 32 On the Edict):

But what if the seller did not know that the slave was a thief and asserted that he was of good character and trustworthy and sold him for a high price? Is he liable ex empto? And I would think that he is (et putem teneri). But he did not know! Yet he ought not to have been so quick to make assertions about something outside his knowledge (sed non debuit facile quae ignorabat adsever- are). Thus, there is not much difference between this seller and one who knows but keeps quiet: for the one who knows ought to warn the buyer that the slave is a thief while this seller ought not to be quick with rash ascriptions.

Also D. 18.1.43 pr. (Florentinus, 8 Institutes):

Things which sellers say to commend their wares do not put them under any obligation provided they are obvious (si palam appareant). As where the seller says the slave is fine-looking or the house is well built. But if he says the slave is literate or has some skill, he ought to answer for those qualities. For by reason of these statements he sells for more.

Here Ulpian is cautious, arguing from the stronger case of fault-based liability represented by the seller who knows he is selling a thief. Florentinus, who may actually have been rather earlier than Ulpian, is more firm: the seller must make good statements which enabled him to sell for a higher price. It is not completely clear whether the �obviousness’ which he says lets a seller off is as to the quality or the nature of the statement as mere blurb. Domum bene aedificatam suggests the latter. It might be difficult to see the quality of the work. But nobody takes any notice of the rubbish in estate agents’ advertisements.

So the seller is liable if he accepts liability. And, by the late classical period, he was taken to accept liability if he asserted the existence of a quality in circumstances calculated to induce reliance and raise the price. But what if he said nothing at all about the quality? Did the natura contractus impose any liability on him, independent of fault?

Now we are concerned with the case in which the seller said nothing about the corn or the chair. The buyer just came and said he wanted twenty measures of �that corn’ or that he would give so much money for �that chair’. And the seller obliged. This is difficult and uncertain. There is virtually no doubt that the texts have been man­aged, in order to give effect to Justinian’s views. If you take an extreme view of the degree of interpolation you can come quite respectably to the conclusion that the classical position, outside the aedilician juris­diction, was caveat emptor: �let the buyer beware’, and �let the buyer insist on express terms’. On the other hand there is some evidence of a doctrine to the effect that goods must conform to the description under which they were sold. It is possible that the owner of the poisoned chickens could have maintained an actio ex empto on the warranty principle while the buyer of the wobbly chair could not. On the ground that poisoned feed is not feed at all while a chair which quickly goes wobbly is nonetheless a chair despite being a bad one. In short, the seller ought ex fide bona to answer for the goods being what he held them out to be. A great deal depends on whether one accepts as genuine what is said about Labeo’s opinion in D.19.1.6.4 (Pomponius, 9 On Sabinus):

If you have sold me a certain container (vas aliquod) and have said that it takes a certain amount or holds a certain weight, I shall sue you ex empto should it prove smaller (si minus praestes). But if you have sold me a container with the assurance that it does not leak (ita ut adfirmares integrum) then, if it does leak, you will be answerable to me even for what I thereby lose. On the other hand if the intention behind the agreement was not that you should provide a sound container (si vero non id actum sit, ut integrum praestes) you must answer only for bad faith [according to Sabinus] (dolum malum dumtaxat praestare te debere). Labeo thinks the opposite. He maintains that the only correct approach is that, unless there is an intention expressed to the contrary, the container must be sound (nisi in contrarium id actum sit, omnimodo integrum praestari debeat). And that is right.

Here the square brackets round Sabinus show that he does not figure in the Latin. There is an accusative and infinitive construction which in commentaries ad Sabinum usually indicates an opinion of Sabinus unless the contrary appears. And then there is the opposite view of Labeo which meets with Ulpian's et est verum. Labeo seems to be saying that in a sale of a vessel or container (vas aliquod) the seller must answer for its not being leaky unless the opposite intention can be inferred from the conduct of the parties, as for instance by the buyer's making clear that he only wanted it for its ornamental value or only to store apples in. Which is as much as to say that, absent such an intention, a vessel must be able to hold liquids in order to come within that description. If you believe that Labeo held that view you can infer that he would have said that corn is also not corn if it is contaminated with poison.

Similar to The Leaky Container is The Not So New Clothes. This text runs into deeper problems as it proceeds, but the first few lines appear to confirm what we have seen ofLabeo's views. D. 18.1.45 (Marcianus, 4 Rules):

In his book called Posteriora, Labeo writes that, in the case in which a man buys, as new, clothes which are actually �interpola', Trebatius's opinion was that so long as the buyer was unaware that the clothes were �interpola' the seller must make good the buyer's interest.

�Interpola' is left in Latin because it is difficult to translate. It may just mean �second-hand'. But more likely it means �made up from cloth cut from discarded clothes'. In which case the nearest analogy in our experience is buying tyres which turn out to be retreads. Either way, the trouble is much the same. The buyer finds worn patches, the life already gone out of the cloth. And again it is Labeo who favours liability.

This is only part of the evidence and, as you will see, even this is not unequivocal. Perhaps the safest summary is conditional. If there was any implied warranty liability in classical law, it was on these lines: the seller was bound to answer for the res being what it purported to be, having the characteristics and qualities definitively necessary for it to qualify for its description. Vessels must hold liquids in order to be vessels, unless proclaimed to be �for dry goods only'; and clothes must be made of new cloth unless proclaimed to be cast-offs or re-makes.

(ii) Fault liability. Obviously, the less warranty liability you are able to believe in, the more important becomes the fault liability. People do not embrace the burden of proving bad faith unless they have to or unless there is some considerable advantage in doing so. If there was no warranty liability, or none except such as was accepted by the seller, then clearly buyers had to turn to fault liability. Even if there was a warranty liability there may have been advantages in proving bad faith. Even in classical law the damages may have been higher, as they certainly were under Justinian. And it may have been possible to reach defects less fundamental in character than those within the warranty liability. Also conformity to good faith was non-excludable.

In the previous section we dealt first with the seller who ascribed qualities to the res and then with the one who said nothing but simply held the res out as having a particular identity or description. The same division can be made here.

The seller who knowingly makes false statements of quality is liable unless they are obvious puffery. This follows a fortiori from what will be said about knowing silence. And we have already seen that in D.19.1.13.3 Ulpian argues towards a warranty liability for innocent statements from the case of statements made in bad faith.¹¹ The objection he puts to his cautious conclusion for liability is â€?But he did not know!' It is obvious that had he known, there would have been no hesitation.

If it is right that only the late classics managed to assimilate misrep­resentations and warranties (i.e. mere statements and statements accept­ing liability for what was said), the only recourse in all the earlier period ^[25] for misrepresentation will have been by proof of fraud. In other words, there will have been no liability for misrepresentation as such. Only a liability for mala fides, one way of incurring which was to make knowingly false statements about the goods being sold.

11

Fraudulent silence is the more interesting case. This happens where the seller knows that the buyer is going to suffer a disappointment in the res but does not speak out to warn him. There is no doubt that this kind of reticence did give rise to liability. But there is great difficulty in making out the scope of the duty of disclosure. The reason is obvious. On an extreme view the duty would oblige the seller to undo all the advantage of his bargain. You never met a shopkeeper who thought he had to tell you that you would get the same thing cheaper and better round the corner. Clearly a line has to be drawn. And it is not at all clear where it was drawn.

The operation of this liability for non-disclosure is illustrated in sales of land when legal defects appear after conveyance. Suppose you buy land from me and find afterwards that there is a right of way over it. You can bring the actio ex empto if you can show that I knew of the adverse servitude and did not tell you. D.19.1.1.1 (Ulpian, 28 On Sabinus):

A seller will not escape the actio ex empto if he has concealed a servitude when he knew it was owed, provided only that the buyer did not know of it. For everything done contrary to bona fides comes into the actio ex empto.

On the other hand if I myself was innocent you can only complain if I accepted liability for the land's being clear of servitudes, which I probably would have done, if at all, by using the conventional phrase �optimus maximusque’. D.18.1.59 (Celsus, 8 Digest):

When you sold farm land, you did not say �ita ut optimus maximusque’ (subject to the term �best and greatest’). It is correct, as Quintus Mucius held, that the land must be made over as it actually is and not free from burdens (non liberum sed qualis esset). And the same applies to urban holdings.

The picture here is this: the buyer’s only hopes of redress lie in an express warranty or, failing that, in fault-based liability. There is no implied warranty-liability.

From a much earlier period Cicero gives an example which has a modern flavour. It concerns the sale of a house subject to a demolition order. Only the reason for the order is antique. Cicero, De Officiis, 3.66-67:

Titus Claudius Centumalus had a house on the Mons Caelius. Its height obstructed the auspices. The augurs wanted to conduct an augury on the citadel and therefore ordered him to demolish the house. He then advertised it for sale and it was bought by Publius Calpurnius Lanarius. The augurs served the same notice on him. Calpurnius demolished the house, but when he discovered that Claudius had advertised it after getting the augurs’ demolition order he took Claudius to law on the formulary issue �whatever he ought to do for him in good faith’. Marcus Cato...pronounced this judgement as iudex: �Since, in the course of selling, the seller knew of this matter and did not reveal it he ought to make good the buyer’s loss.’ Hence his opinion was that good faith required that the buyer should be told of a vitium known to the seller.

How far did this liability for mala fide non-disclosure go? That is, what limits were put on it to stop the erosion of every advantage in a bargain? There seem to be three possibilities. First, that it applied only to legal defects in land. Second, that it applied within the actio ex empto to the same defects as were covered, within the aediles’ jurisdiction, by the aedilician edict. Third, that it applied to all cases in which, according to prevailing commercial morality, suppression of the truth would have been regarded as immoral. This third possibility is loosest and leaves most to the judge, but I incline to the view that it may be the right answer. We have seen, in the case of The Leaking Container, that Sabinus thought, contrary to Labeo, that the seller should be liable only for dolus malus, and one can infer that he meant that the seller would be liable if he knowingly sold a leaky container to a customer who expected one which did not leak. If that is what he thought the first and second possibilities must be wrong. On the other hand it is difficult to imagine Sabinus saying that the seller of a not very well made chair must explain to his buyer that it would probably not last very long. This is much spongier that it should be. But undeniably the position is unclear.

(iii) An overview. This is by way of summary of the last two parts, making the picture clearer than it really is. How should the dissatisfied buyer phase his inquiry? Perhaps in this way. First, did the seller ascribe qualities to the res? If so, he may have warranted or merely represented. If he warranted, he is liable without fault. If he merely represented, he will be liable if he knew that his statement was untrue (unless it was obvious puffery); and towards the end of the classical period he became liable as though he had warranted, irrespective of fault. Because war­ranties and misrepresentations were assimilated. Until then the inno­cent misrepresentation triggered no liability as such, i.e. unless the seller would have been liable even if he had not uttered it. Second, if the seller ascribed no qualities to the thing, he may yet be liable in one of two ways. Either he may fall within the doctrine of dishonest silence, incurring fault-liability for dishonest non-disclosure, or he may fall within an implied warranty liability for defects so serious as to deprive the res of the identity under which it purported to be sold. Both these doctrines are of uncertain scope. Dishonest silence certainly applies to legal defects in land and may find its limit merely in prevailing notions of commercial morality. The implied warranty applied only to defects essential to the description of the thing, few of which would not be patent.

(iv) The aediles' edict. This edict supported an action superintended by the aediles themselves, not by the praetor. Its life was thus separate from that of the actio ex empto.

The scheme of the edict was to impose a warranty liability for undeclared defects. For slaves, the seller was to declare all disease or defect (morbum and vitium) and certain listed non-physical vices (fugi­tive, wanderer, noxally liable, capital offences, sent into arena, suicide attempts). Then, if the slave turned out contrary to what had been declared, or to what had been dictum promissumve in circumstances such that the dictum or promissum ought to be made good, two remedies were available. Within six months, an action for redhibition, a giving and taking back; and within twelve an action for reduction of the price. Then, in relation to beasts of burden (iumenta), similar provisions for disease and defect. The edict is said to be designed to counter fraud. But liability under it is not a fault liability. The plaintiff buyer only has to prove that the res was not as declared. So fraud was countered by giving a remedy against fraudulent and innocent alike. Patent defects and very trivial defects were ignored. For slaves the defects within the edict were those which interfered with his work and service (usus et ministerium).

(v) Justinian's law. He merged the actio ex empto and the aediles' edict into one body of law, making the aedilician principles apply throughout, even to property other than the slaves and animals originally contemplated. This was achieved by interpolation. Which means that we cannot easily see what, if anything, had already been done by his time. It is this interference in the texts which makes the whole topic so difficult. It may yet be demonstrable that the absorption of aedilician principles into the actio ex empto was largely achieved well before Justinian.

The effects of the merger were to introduce a general liability for dicta other than puffs. On the remedial side, the aediles' actions for redhibition and reduction of price were made available alongside the normal damages under the actio ex empto, subject to this innovation. Normal damages covering the buyer's whole loss were retained for the case in which the seller had been guilty of bad faith while redhibition and reduction were made available against the innocent seller. In this way a distinct advantage developed for those who were willing to prove bad faith.

It is often said that the law relating to latent defects is easier to understand if one starts from Justinian and works backwards. It is then a question of disentangling the strands which contributed to the general warranty. On the one hand the warranty liability under the aediles' edict; on the other the limited warranty liability under the actio ex empto and the liability for mala fides under the same action. The confluence of these streams will remain the pattern of the historical development even if it turns out that they met a good deal earlier than in Justinian's time.

iii. The action against the buyer

The seller's action is the actio ex vendito. Theformula is the same with the parties turned round and �sold' put in the demonstratio instead of �bought'. And of course the facts to be proved to substantiate the demonstratio would be the same.

Emptio-venditio is bilateral, and we have seen that it was differences between the obligations of the two parties which forced permutatio (barter) into a separate compartment. What could the seller sue for? From the other side, what were the obligations of the buyer? The buyer's side is more straightforward, chiefly no doubt because the problem of latent defects is not found in relation to money.

The actio ex vendito lies above all for the price, which the seller can claim so long as he himself has performed or is ready and willing to do so. Whereas the seller is only obliged to warrant against eviction, the buyer must make the seller the owner of the coins. Hence if he pays over money belonging to a third party he is immediately in breach, even before the third party himself takes action.

Though there must be money in the price in order to identify the buyer there may be other elements too. And the seller's action lies for them. D.19.1.6.1,2 (Pomponius, 9 On Sabinus):

1. If I have sold you an apartment block for a fixed sum of money and an undertaking to repair another block of mine, I will have the actio ex vendito to enforce the repair. But if the agreement charged you only with the repair it would not amount to emptio-venditio, as Neratius also wrote. 2. But if I have sold and conveyed a plot of land to you for a fixed price on the terms that you are to build a block on it and then re-convey a half part to me, it is right that I can bring the actio ex vendito both to enforce the work of building and to get you to execute the reconveyance of the block once built. For it is clear that so long as there remains with you something attributable to the res vendita I retain my actio ex vendito.

Though more various than simple payment of coins, all this can be summed up in the one duty to pay the price.

The buyer's other principal obligation is to take delivery. The seller must make traditio or, originally and where necessary, mancipatio. But that does not mean that he has to seek out the buyer or send the res to him. The terms of the contract may impose that obligation but if they do not then the buyer must come to take the res from the seller. D.19.1.9 (Pomponius, 20 On Sabinus):

If someone who has bought stones from an estate will not take them the actio ex vendito can be brought against him to make him remove them.

See also D.18.6.1.3 (Ulpian, 28 On Sabinus):

The seller of wine can even pour it away if he specified and made available a time for it to be measured out and it was not measured out. But he will not be able to pour it out at once, before giving the buyer notice before witnesses to fetch the wine away or leave it in the knowledge that it will be poured off. However if having the right to pour it away he abstains from doing so, that is praiseworthy and for that reason he can demand payment for the vessels, so long at least as he had an interest in their being emptied of the wine in them, as for instance if he was going to hire them out or needed to hire in others for himself...

iv. Risk (periculum)

The Latin periculum is more familiar as �danger’. We are all exposed to danger all of the time. That is what insurance is about. And social security too. Leaving those safety schemes aside, we can say that if disasters do materialise there either will or will not be someone to sue. If you drive into my car negligently I can sue you for compensation and thus shift the loss to you. If my car is swept away in a flood or overturned by a whirlwind there will almost certainly be nobody to complain against. So the disaster will be all my own. Just like catching measles.

The dangers to which property is exposed are destruction (including loss) and damage (including deterioration). It is convenient to divide these even though in practice it is sometimes difficult to say whether the damage is so bad as to count as destruction. If property never changed hands the danger of damage and destruction would simply be borne by the owner. Sale does involve a change of hands. So there is obviously a question to be faced. In the story of a sale when exactly does the burden of bearing these risks pass from the seller to the buyer? That is the question when. But there is also a question what: What is involved in the passing of the risk?

A. What is involved? It is much easier to start by concentrating solely on destruction. Think of a horse which dies or a vase which smashes into pieces. Once the risk of destruction has passed to the buyer he falls into the same position as an owner has in relation to goods which he has no notion of selling. And that is true even though he, the buyer, may not yet be owner. For it is not to be assumed that risk and title pass together. So once the risk has passed the buyer has to ask himself what if anything he can do about the loss which he has suffered. Can he sue anyone? Was it through someone’s want of care that the horse died or was killed, or the vase shattered? He may be able to maintain a claim or he may not. What he cannot do once the risk has passed is to shift the loss on to the seller by calling the contract off and refusing to pay the price or, if it has been paid, demanding it back.

Suppose the poor horse is struck by lightning a year after I bought him. It is obvious that I cannot go to the seller and say I want the sale reversed. And equally obviously if I still have not paid I must pay. And the same is true from the very moment after the risk passes. The year only makes the point seem too plain for argument. The sooner the destruction happens the more it seems as though the buyer has to pay for nothing. And if the risk passes before he obtained possession it may also materialise before he obtains possession. In which case he must pay even though he gets nothing at all, in the sense at least that he gets not even one second's enjoyment. Suppose I buy the horse which from your window we can see standing in your field. Let it be given that the risk of destruction passes. As we walk out to the field a storm breaks and the horse is struck. I must pay. The modern moral is that the time from which to insure is the moment at which the risk passes.

All this can be summed up thus. If destruction happens after the risk of destruction has passed, the buyer cannot shift the loss back to the seller by refusing or reclaiming the price. His only hope of shifting the loss is that the circumstances may be such as to make someone, perhaps even the seller, liable for the disaster.

What if the destruction occurs before the risk passes? Then the seller cannot shift the loss to the buyer by claiming or retaining the price. It is his affair, just as though there had been no sale in view. It is strictly speaking a separate question whether he might, in addition to bearing the loss of his res, also have to pay damages for non-delivery. The answer is that he will not, because the contract will be aborted by the disappearance of its subject-matter. The only doubt is whether that is an invariable rule or a rule of construction which holds good in 999 cases out of a thousand.

Damage is rather more difficult. Once again the best approach is to start by asking about damage which happens after the risk passes. Suppose the horse is lamed or the vase chipped. The seller can shrug his shoulders. It is none of his business, unless it happens that it was his fault. Suppose that the buyer has insisted on a term in the contract that the res shall be in perfect condition. It does not make any difference, because, whatever the contractual standard, deteriorations happening after the risk passes are the buyer's affair. If the vase was perfect when the risk passed the loss from the chip must be approached in the same way whether it happens one second or a year later.

What about damage before the risk passes? This is where there is a problem. Suppose a system in which there is no implied liability for quality and a contract in which there is no express term about it. Here damage which happens �before the risk passes' is a contradiction in terms. For the quality of the thing is at the buyer's risk throughout. Let him insert a contractual standard. Say, specifically that the horse shall not be lame. Now the notion of risk of damage, or at least of risk of lameness, begins to make sense. The contractual standard requires no lameness, and the lameness has supervened before the risk passed to the buyer. So the seller will be liable for breach of the standard laid down in the term against lameness. There is no ground for suggesting that the sale is void because of the lameness, though that result could be reached by the use of conditions.

The thrust of the last paragraph is that, for damage, the notion of a transfer of risk is a virtually meaningless concept. To gain meaning it requires �damage’ to be understood as �deterioration below the con­tractual standard set by terms as to quality’. If there is no such standard there is no time during which, once a sale has been agreed, it can be said that the risk of deterioration was on the seller. And contrary-wise, to affirm that the risk of damage is on the seller is to affirm that there is indeed a contractual standard in relation to quality, set by express, implied or imposed terms.

B. When does the risk pass? The first rule is that this is something within the power of the parties to choose. Failing express agreement, the risk passes on conclusion of the contract. J.3.23.3:

When emptio-venditio is contracted (which we have described as happening as soon as there is agreement on the price when the business is done without writing) the risk in the thing sold immediately attaches to the buyer, even if the thing still has not been delivered to the buyer.

So the risk passes where the agreement is complete, not when the res is delivered. Hence, before ownership. The maxim res perit domino (when property is lost, it is the owner who suffers) does not apply. The reason must be that it is from the moment of being bound to give the thing up that the owner loses his economic interest in it. From then its value and potential belong to the buyer, whatever the technical location of title.

This statement in the Institutes is a simplification. It covers the common case. More accurately it should be said that the risk passes at the point of â€?perfection’, which can come later than conclusion of the contract. If there is a condition to be fulfilled or if there needs to be weighing or measuring to identify the res (20 gallons from this vat) or to ascertain the price in figures (all this vat at £20 per gallon), the contract is binding but â€?imperfect’.

In such cases, where there is an imperfect phase between contract and perfection, there is sometimes room for an argument that the sale can become perfect even though the res is destroyed. For example, you buy this foal subject to the condition that its mother wins tomorrow’s race. The foal dies and the mother wins. Or you buy all the wine in this vessel at £1.00 per litre. The vessel is punctured and the wine spills out, but I mend the hold and, filling the vessel with water, measure the contents at 10 litres. This argument leads to the suggestion that the risk of destruction might pass retroactively from the notional perfection back to the making of the contract, so that you would have to pay. But that argument did not prevail. The sale was avoided with the destruc­tion of the res. However, the contrary position appears to have been taken in relation to deterioration in the imperfect phase. But this is not certain.

In all cases in which traditio does not follow instantly upon perfection of the contract, there is a phase during which the buyer bears the risk though the seller is in possession. It should not be thought that the buyer is unduly vulnerable during that time, for the seller is subject to a high standard of care and will be liable to the buyer for losses inflicted for want of that care. Also, an entirely different kind of mitigation of his difficulty gives him an entitlement to fruits and other enhancements which accrue during the period when he bears the risk. J.3.23.3 (continued):

And so if the slave has died or been harmed in some part of his body, or the house has been wholly or partly consumed by fire, or the land has been wholly or partly swept away by a torrent or has become worse or smaller by far through flood or the destruction of trees by a gale, the loss lies on the buyer. He will have to pay the price even if he does not get the thing. For whatever happens without deceit or fault (sine dolo aut culpa) on the part of the seller, as regards that the seller will be safe. But if after the sale anything accedes to the land by fluvial accretion (per alluvionem) that advantage goes to the buyer. For whoever bears the risk of disaster should also enjoy the hope of advantage (nam et commodum eius esse debet cuius periculum est).

v. The passing of property

The outline is as follows. First, the agreement itself did not transfer ownership: contract and conveyance were separate events. Second, sales of res mancipi required mancipation or cession in iure if dominium was to pass. Otherwise the buyer would just be put in via usucapiendi, with praetorian protection till dominium accrued by usucapio. Third, for res nec mancipi, and in Justinian’s law for res of all kinds, traditio was necessary. In classical law the traditio following sale was sufficient in itself, without payment of price. G.2.20:

And so if I deliver to you some clothes or gold or silver on the basis of sale or gift or some other such cause, the res immediately becomes yours, if only I myself am owner to start with.

However, Justinian introduced a new rule, which he said was an old one. J.2.1.41:

If res are delivered by traditio on the basis of gift or dowry or any other cause there is no doubt that they are alienated. But things sold and delivered by traditio do not become the property of the recipient buyer unless he has paid the price to the seller or given him some other satisfaction as by surety or pledge. And this is laid down even by the Twelve Tables, though it is right to say that it is also part of the ius gentium (i.e. of the ius naturale). But it is also true that if the seller grants credit to the buyer the res immediately passes to the buyer.

This same rule is stated in the Digest, attributed to Pomponius at D.18.1.19.^[26] But at D.18.1.53,^[27] Gaius is made to say that either payment or security is necessary if title is to pass, with no mention of the granting of credit. There seem to have been last minute vacillations on the part of Justinian’s commissioners. It is not clear what they were aiming at. The old rule, if it existed, applied to mancipatio. The provi­sion about the granting of credit comes very near to short-circuiting the whole change, since there must be few cases in which the price is not paid and credit is not given. But the effect of the rule is to give the seller the security of continuing ownership in the one case in which he had made no arrangements of that kind with the buyer.

Ad hoc reservation of title by way of security for payment can be troublesome. Subsequent buyers are in danger of getting no title. It may have been considered more convenient to reserve title in every case, thus putting all subsequent buyers on notice to ask for proof of price paid or some arrangement made. But this is a guess.