LEGAL TRANSACTIONS AND THE LEGAL OUGHT
Universities of Maastricht and Hasselt
In 1964 John Searle offered a derivation of an ought-judgment from is-premises in an argument that ran ahead of much of his later work on the construction of social reality. The argument was based on an analysis of promising and as an example this was not the most happy one, because the argument seemed to presuppose the premise that promises ought to be kept.
This paper consists of two parts. In the first part it is shown how the main ideas of Searle’s original argument can be presented in a way that is less vulnerable to the objection that an ought-premise is presupposed. To that purpose the argument is based on an analysis of legal transactions, rather than promises.
The second part of the paper is devoted to the objection that the derived ought is not a ‘real’ ought, but merely a legal ought. It is argued that this objection is mistaken, but the analysis of deontic concepts that is given there also makes it clear that the argument based on legal transactions does not provide the desired derivation of ought from is either. However, what may be considered to be Searle’s main point, namely that the presence of deontic facts can be based on non-deontic ones, turns out to withstand this criticism.
It is widely accepted amongst moral, political and legal philosophers that it is not possible to deduce an ought-conclusion from solely is-premises. That this is impossible is sometimes called ‘Hume’s law’ because of the seeming allusion to it in Hume’s Treatise of Human Nature. The alleged fallacy to make this derivation goes under the name of the (deductive version of the) naturalistic fallacy. In an early paper, John Searle made an attempt to show that it is nevertheless possible to derive ‘ought’ from ‘is’. The derivation went as follows:
1. Jones uttered the words ‘I hereby promise to pay you, Smith, five dollars’.
2. Jones promised to pay Smith five dollars.
3. Jones placed himself under (undertook) an obligation to pay Smith five dollars.
4. Jones is under an obligation to pay Smith five dollars.
5. Jones ought to pay Smith five dollars.
Searle argued that the relation between a statement in this list and its successor is either an entailment or at least not contingent, and moreover that the relation could, where needed, be made into an entailment by the addition of a premise which was neither an evaluative statement, nor a moral principle, nor anything of the sort.
Searle’s argument is less than convincing, but nevertheless correct in its underlying idea. It is less than convincing because of the presupposition needed to get from (2) to (5), which is presumably something like:
For any x and any A, if x promised to do A, then x ought to do A.
Although some caveats may be necessary to account for exceptions and cancelling conditions, something like this presupposition is necessary to make the argument deductively valid. The problem is that this presupposition is a sentence which expresses an ought. Searle tries to tackle this objection by the claim that the sentences which make the transition between the sentences of the argument deductively valid are analytic (literally: ‘tautologies’), but even if Searle is correct in this claim, the issue remains that analytic ought-sentences are also ought-sentences. Therefore, Searle did not succeed in deriving ‘ought’ from ‘is’.
Nevertheless, Searle’s argument that it is possible to derive ‘ought’ from ‘is’ has an underlying idea which is correct. This underlying idea is that some facts, including some institutional facts, lead to new obligations and therefore new oughts, and that the presence of these facts itself does not depend on obligations or oughts. In this paper I will first try to illustrate the correctness of this idea by using contracts as example, instead of promises. It is possible to create obligations by means of contracts, and it is possible to describe this creation in terms of a valid argument with only is-premises and as conclusion that the contract partners are under an obligation to do something. The step from the presence of a legal obligation to the presence of a legal ought then seems to be minor. Subsequently I will discuss an objection against this derivation, namely that the derived ‘ought’ is merely a ‘legal ought’ and not a full-blown ‘ought’. This discussion will also shed light on the last step of the derivation, namely from the premise that somebody is under an obligation to do something to the conclusion that this person ought to do this.
Contracts are a means by which legal subjects can change the legal positions of themselves or other persons. One fruitful way to look at legal transactions in general and contracts in particular is to see them as intentional changes to the ‘world of law’.
We are all familiar with the physical world. It consists of a large number of facts. The facts in the physical world obtain to a large extent independent of human beings. The social world, or social reality, does not only depend on what is physically the case, but also - and to a large extent - on what people believe the social world is. A fact in the social world can obtain because sufficiently many members of a social group believe that it obtains because people believe so, and also believe that (sufficiently many) other members of the group have the same belief, both about this fact and about what the others believe. Jane may, for example, be the leader of an informal group, because most members of the group take her to be the leader and believe that the others take her to be the leader too and believe that the other members do the same.
In modern societies, however, many facts obtain because of the operation of rules, including legal rules. These rules deal with how people should behave towards each other, but also with the proper use of language, with the definitions of games, and with the membership of socially defined sets, such as the set of legal rules (e.g Hart’s rules of recognition). If the conditions of these rules are satisfied, their consequences hold in social reality. The part of social reality that is the result of the application of rules is called the institutionalised part of social reality. Typical phenomena within the institutionalised part of the social world are the existence of money, promises, the law and everything created through the law, such as officials, legally defined organisations and most legal rules. The world of law is part of the social world and in particular the institutionalised part of it. The existence of large parts of the law, both rules and legal positions, is based on the operation of rules which attach legal consequences to events, including legal transactions.
A legal transaction, such as a contract, takes place, if somebody performs an act with the intention to create particular legal consequences (changes in the world of law), and the law attaches the intended legal consequences to this act precisely because they were intended.
Searle made a number of distinctions between types of speech acts which are useful for a better understanding of legal transactions. An auxiliary distinction in this connection is that between directions of fit. It is illustrated by the following example. Suppose I make a shopping list that I use in the supermarket to put items in my trolley. A detective follows me and makes a list of everything that I put in my trolley. After I am finished, the list of the detective will be identical to my shopping list. However, the lists had different functions. If I use the list correctly, I place exactly those items in my trolley that are indicated on the list. My behaviour is adapted to what is on my list. In the case of the detective it is just the other way round; the detective’s list reflects my shopping behaviour. If we consider my behaviour as (part of) the world, we can say that my shopping list has the world-to-word direction of fit, because my behaviour (the world) must fit the words on the list (the words). The detective's list, on the contrary, has the word-to-world direction of fit, because his list must fit my behaviour.
The direction of fit holds between the propositional content of a speech act and the world. By means of the distinction between directions of fit, it is possible to distinguish between kinds of speech acts:
Assertives express that something is the case. For instance, the sentence ‘It's raining’ can be used for an assertive speech act, and then the assertion expresses the state of affairs that it is raining. Assertives have the word-to-world direction of fit; they are successful if they are true.
Directives are attempts of the speaker to get the hearer to do something. For instance, the sentence ‘Give me your money’ can be used for a directive speech act. Directives have the world-to-word direction of fit, and are successful if the hearer does what he was directed to do. Because the success of a directive depends on subsequent compliance, I will call the world-to-word direction of fit of directives indirect.
Constitutives bring about changes in the world, in particular in the institutionalised part of social reality. Examples are acts of baptising, appointments in functions, and – as will be discussed later – promises and legal transactions. Constitutives also have the world-to-word direction of fit, but in a very different way than directives. If a constitutive is valid, its propositional content becomes true in social reality.
The distinction between directives and constitutives may be illustrated by the difference between orders and commands, where both terms are used in a technical sense. An order will normally exercise some psychological pressure on the hearer to do what he is being directed to do. However, there is no guarantee that the order will be obeyed and that the world will actually come to correspond to the directive's propositional content. That is why Searle writes about the fit of successful directives, and ‘successful’ means in this context effective. Orders do not lead to obligations, and orders are not valid or invalid, because giving orders does not presuppose a setting of rules.
Everybody can order anybody, but a command requires a setting in which the commanding person has some authority over the person who is commanded. Moreover, a successful command brings about that the person who was commanded is under an obligation to do (ought to do) what was commanded. A command is successful if this obligation is brought about. It is not required that the resulting obligation is also acted upon. Therefore the success condition of a command is that the command is valid, not that it is effective. The possibility that a command is valid presupposes a setting of rules that gives the commander the power to create obligations through the orders and which obligates the commanded person to do what was commanded.
In general, constitutives need to be successful to create the world-to-word fit, but their success is not the effectiveness but rather the validity of the speech act. Searle correctly remarks that constitutives (he calls them ‘declarations’) normally require an extra-linguistic institution, a system of constitutive rules, in order that the declaration may successfully be performed. For instance, there are rules that lay down how the appointment of a chairwoman should take place. If these rules are followed in a concrete case, the appointment in question is valid, and the appointed woman has become the chair. The institution not only defines when constitutive acts are valid, but also connects consequences to valid constitutives, for instance that somebody has become the chair. These consequences are changes in the institutionalised part of the world, which account for the world-to-word fit of constitutives. Because a successful constitutive ‘automatically’ leads to the realisation of its propositional content, I will call the world-to-word direction of fit direct (as opposed to the indirect fit of directives).
Promises are constitutives by means of which obligations are created. In this respect they are like commands. The difference between the two is that a command creates an obligation for the commanded person, not for the commander, while a promise creates an obligation for the promisor.
Legal transactions can very well be taken as a special kind of constitutive acts. They have the direct world-to-word direction of fit, meaning that if a legal transaction is valid, its propositional content becomes true in the world of law as a consequence of the transaction. Just as with other constitutives, legal transactions can only work within a setting of rules. These rules define amongst others how legal transactions are to be performed, which persons are competent to bring about particular legal consequences, who have the capacity to perform legal transactions and which legal consequences are connected to the successful performance of a legal transaction. The very idea of legal transactions is that these legal consequences are by and large the ones intended by the performer of the transaction. Therefore a legal system can be said to acknowledge legal transactions if it approximately attaches the intended consequences to legal transactions for the reasons that they were intended.
Obvious as this view of legal transactions may seem at first sight, its implications are far-reaching. Amongst others it means that it is – almost - possible to derive an ought-conclusion from is-premises only. This will be illustrated by means of an analysis of contracts.
The only point of making promises is to undertake obligations. Although it is possible to undertake obligations by means of contracts as well, contracts can be also used for other purposes that cannot be achieved by promises. It is for instance possible to appoint by means of a contract an arbiter who is empowered to decide over conflicts that might arise in connection with the execution of (the rest of) the contract. Although it is possible to construct such an appointment as undertaking the obligation to do what the arbiter decides, this would misrepresent the nature of such an appointment. By means of a contract, the parties create or abolish facts in the world of law, to the extent that they indicate so in the contract, and to the extent that they are empowered to do so.
Although contracts do not necessarily lead to obligations – they may for instance be confined to the cancellation of obligations, or to the division of risks in case of default in an already existing contract - they often do. Surprisingly, few people see this as involving some variety of the naturalistic fallacy. It might be objected that contracts only seemingly bridge the gap between is and ought because the obligation to do what was contracted is based on the rule that contracts ought to be complied with. The contract itself would on this view be nothing more than a specification for a concrete situation of what this general obligation implies. It is, however, questionable whether the rule exists that contracts ought to be complied with. The point of contracts is more general than merely that contracts facilitate the intentional creation of obligations. Their point is that the facts established by means of the contract hold between the contract parties. Underlying contracts is not the rule that contracts ought to be complied with, but the rule that the facts which a contracts aims to bring about, actually come to existence if, and to the extent that, the contract is valid.
The rule that what parties agreed to holds between the parties itself does not impose any obligations. That obligations result from most contracts is because by means of most contracts the contract parties create obligations between themselves. Notice the emphasis on ‘create’. The obligations were not yet there before the contract; they are the result of the contract. The presumed rule that one ought to obey one’s contracts is superfluous. If the contract does not create obligations, but aims for instance to cancel existing obligations, there is nothing to obey, and the rule would not make sense. If the contract does create obligations, the rule that one ought to comply with these obligations would effectively be that one ought to do what one is under an obligation to do. That would be an analytical rule which does not make much sense either. So there is no role for the rule that contracts ought to be obeyed. The obligation to do what one contracted to do therefore does not derive from such a rule. The obligation is created by means of the contract and it is a new obligation that did not yet exist before the contract, not even in the more abstract form of an obligation to comply with one’s contracts.
The same point can also be made in a slightly different way. The operation of contracts by means of which obligations are created can be analysed in two ways. According to the first way, there exists a prior obligation to do what one has contracted to do. This obligation already existed, and the sole function of the contract is to give this general obligation a specific content by indicating exactly what parties have undertaken to do. This style of analysis can also be used for the analysis of promises, if one assumes that the practice of promises includes the existence of a prior obligation to keep ones promises. The function of actual promises is then to provide content to this general pre-existing obligation. Logically this analysis boils down to something like the following:
I. For all x it holds that if x has contracted/promised to do A, then x is under an obligation to do A.
II. Jones has contracted/promised to pay Smith five dollars.
III. Therefore: Jones is under an obligation to pay Smith five dollars.
IV. For all x and all A it holds that if x is under an obligation to do A, then pro tanto x ought to do A.
V. Therefore: Jones ought to pay Smith five dollars.
Notice that the first premise, which expresses the pre-existing general obligation, is an ought-premise, or – better - a premise that expresses a general obligation which in turn pro tanto implies an ought. The ought-conclusion is on this analysis not derived from is-premises only. Moreover, the ought of the conclusion derives from the ought of premise I. As we will see in section 5.4, this interpretation of the first premise is misguided.
The second way to analyse the operation of contracts emphasises that contracts, like other legal transactions, bring about the facts in the world of law which the parties intended to bring about and which they expressed in the propositional content of the contract. On this analysis there is no pre-existing obligation, and the contract can therefore not be interpreted as a specification of an obligation that already existed. Instead of the obligation, there is a pre-existing rule which makes legal transactions possible. In the case of contracts, this rule approximately holds that the facts which the parties intended to bring about will actually hold in the world of law. Notice that there is nothing ought-like in this rule; the rule equally holds for cancellations of duties, for the empowerment of arbiters, for the foundation of organisations, as well as for the creation of obligations. Logically this analysis boils down to something like the following:
VI. For all states of affairs *s it holds that if a valid legal transaction was performed with as content [state of affairs *s is the case] , then, from that moment on, state of affairs *s is the case in the world of law.
VII. Jones has concluded a valid contract with Smith with, amongst others, as content [Jones will be under an obligation to pay Smith five dollars].
VIII. Therefore: after the contract, Jones is under an obligation to pay Smith five dollars.
IX. For all x and all A it holds that if x is under an obligation to do A, then pro tanto x ought to do A.
X. Therefore: Jones ought to pay Smith five dollars.
Notice that the counterpart in this analysis of ought-premise I of the earlier argument is premise VI, which does not mention an obligation. Premise VII may seem to contain an (implicit) obligation, but it does not express an existing obligation, but is merely part of the propositional content of the speech act by means of which the obligation was created. It is, in a sense, an ought between quotes. To use a memorable phrase of Quine: the ‘ought’ is mentioned, not used. Premise IX, however, does contain an ought, since it expresses that obligations lead (pro tanto) to oughts. So the derivation on the second analysis turns out to be an example of a derivation of an obligation from a set of is-premises, but not of the derivation of an ought-conclusion from is-premises.
Even on the second analysis, Searle’s derivation turns out not to be a derivation of ought from is. In section 5.7 we will return to the impact of this observation, but running ahead of of that discussion, let us grant Searle that the step from the existence of an obligation to an ought-conclusion is innocent. Moreover, it is quite plausible that the important step for Searle was the derivation of the obligation from is-premises, rather than the additional step from the existence of the obligation to an ought-conclusion.
The above argument how legal transactions make it almost possible to derive ought-conclusions from purely is-premises may be the target of the objection that the ‘ought’ of conclusion X is not a real ought but only a legal ought, or to use the phrase coined by Hare, an ‘inverted comma’s ought’. Jones only ought to pay Smith five dollars to the extent that Jones intends or ought to obey the law and the obligations imposed by it. To deal with this objection we must make a more detailed analysis of what an ought is, and how it relates to (other) deontic notions such as obligation, duty and should. This analysis will not only show that the first analysis of Searle’s argument is fundamentally mistaken, but also why it is not possible to derive an ought-conclusion from is-premises, and why this impossibility is not so troublesome for Searle’s argument as might seem at first sight.
To get a clear view of what this objection amounts to, we must first distinguish between guiding reasons and motivating reasons. This distinction is necessary to refute the idea that a legal ought only obliges those persons who are motivated to comply with the law. Motivating reasons are beliefs which tend to motivate the person who holds them to act in a particular way. An example would be the belief that it is raining, which tends to motivate John to carry an umbrella. Which beliefs are motivating reasons depends on the psychological make up of the person who holds these beliefs. For this reason, a motivating reason is always person-relative, a reason for a particular person.
Motivating reasons must be distinguished from guiding reasons. Guiding reasons are normally not beliefs but facts, and they are facts which make that a person ought to act in a particular way. An example would be the fact that it is raining, which is a (guiding) reason why John ought to carry an umbrella. If it is raining, but John believes the opposite, John has a reason to carry an umbrella, but will normally not be motivated to do so. However, if it is not raining, but John believes it is, John will be motivated to do what he has no (guiding) reason to do.
The objection that a legal ought is not a real ought may be inspired by the idea that an ought which does not motivate is not real. If this is the case, the objection is misguided, because what one ought to do legally depends on guiding reasons, not on motivating reasons. It is possible that a person has guiding reasons which make that (s)he ought to do something, and that this person is nevertheless not motivated to act accordingly. It may, for instance, be the case that a convicted criminal legally ought to turn himself in, but is not at all motivated to do so.
There are, however, different ways to interpret the objection. One is that a real ought is what one ought to do all things considered, while a legal ought is confined to a single point of view, namely the legal one. The other interpretation is that there is no guarantee that the standards which determine the legal ought are rational ones. Maybe it is irrational to comply with the law, not only because there may be other reasons that need to be taken into account, but also because the legal reasons which are irrational themselves are therefore no ‘real’ reasons at all. It may, to stick to the example, not be rational for the convicted criminal to turn himself in. He is not only not motivated to do so, but rationally – one might argue- he should not turn himself in either.
Both objections suffer from a misunderstanding of what an ought is. An ought is not necessarily connected to action, but rather to a finite set of standards. This has three consequences, namely that
1. there can be an ought relative to standards to which one has not committed oneself, a ‘detached ought’;
2. the idea that a real ought is necessarily an ought all-things-considered is misguided;
3. it is possible that one ought to do something, although it is not rational to act accordingly .
An indication that the connection between ought and action is less close than is often assumed is that many oughts have little to do with action: the box in which my cellular phone was sent to me ought also to contain a user guide; according to its index, a book ought to contain 250 pages; a proper bicycling network ought to be supported with sufficient road signs; software ought to have a good user interface … etc. Although it is possible to construct an action-oriented ought in all of these cases, by assuming that somebody ought to see to it that what ought to be the case is also actually the case, this ought-to-do is not expressed by the example sentences. Moreover, it is doubtful whether every time where the actual situation falls short of what it ought to be, some particular person ought to correct this. It turns out that there are at least two kinds of ought. On the one hand we have an ought-to-do, which is connected to what a person has reason to do. On the other hand there is an ought-to-be, which expresses how the facts should lay without necessarily implying that somebody in particular ought to make the facts comply with this ideal.
However, more important than the difference between the ought-to-do and the ought-to-be, is what these two kinds of ought have in common. In both cases the ought presupposes a standard or a set of standards and it expresses that these standards allow only one option. For instance, if a book has an index indicating how many pages the book counts, then the standard for proper book indexes does not allow that the book counts a different number of pages than that indicated by the index, or that the index indicates a different number than the actual page number. Similarly, if Jones contracted to pay Smith five dollars, the standards for contracts will normally leave only one option open for Jones, namely to pay Smith five dollars. In general, an ought-to-do exists if the standards for proper behaviour only allow one course of action.
A person P ought to do A, if A is the only course of action left open by a set of standards. These standards can be legal, moral, those of etiquette, or those of self-interest. They may even be standards that do not belong to a particular point of view such as goals that happen to be adopted by a particular person, or standards that are derived from several points of view. What is not possible, however, is an ought that is not based on any standard. If we do not have standards to apply, there are no normative constraints on our actions, and it cannot be the case that only one course of action is left open. Therefore, without standards, there is no ought.
A direct consequence of this is that every ought is relative to a set of standards. An action ought legally to be performed if performing it is the only course of action that is compatible with the standards of law (a particular legal system, if you wish). An action ought morally to be omitted if performing this action is not compatible with the constraints imposed by the standards of morality.
Another consequence is that we can determine what we ought to do morally or legally, or relative to any other set of standards, without committing ourselves to act in accordance with these standards. An ought is the outcome of the constraints posed by a set of standards. It is not necessary that the person who arrives at the ought-judgement must be committed to these standards. The conclusion that something ought to be done relative to a particular set of standards does not need to have any consequences for what one is motivated to do. An ought can be completely ‘detached’. This is a consequence of the distinction between guiding reasons and motivating reasons. Guiding reasons, based on standards, determine what one ought to do. Motivating reasons are mental states that cause motivation. It is not so that if a person is not motivated by the awareness of some fact, this fact is not a guiding reason. That the convicted criminal is not motivated to turn himself in does not imply that there is no guiding reason why he ought to do so nevertheless.
Like other ought-judgements, legal ought-judgements can be detached too. Therefore, the possibility that somebody legally ought to do something without being motivated to comply, is no evidence that a legal ought is not a ‘real’ ought.
A possible source of confusion concerning the nature of the ought, especially the ought-to-do, is that this ought is not well distinguished from obligations and duties. If somebody ought to do something, this does not imply that this person is under an obligation to do so, or even that he is under a duty.
Legal and moral obligations are the result of events to which the law or morality connects the existence of obligations. Obligations belong to the institutionalized part of social reality. Typical examples are contractual obligations and their moral counterparts, obligations to fulfil one’s promises, and obligations to compensate for damages which one has caused.
These obligations belong to the standards that determine what ought legally or morally to be done. That is why it is possible to conclude from the fact that Jones is under an obligation to pay Smith five dollars that (pro tanto) Jones ought to pay Smith five dollars. The obligation is a standard that contributes to the ought, but is not an ought itself.
Arguably, the step from the obligation to the ought has a duty as an intermediate conclusion. Because Jones is under an obligation to pay Smith five dollars, Jones has the duty to do so, and this duty is the standard for Jones’ behaviour and in that sense contributes to what Jones ought to do.
Although duties may be involved in obligations, duties are not the same as obligations. Where obligations depend for their existence on events such as contracts, which call the obligations to life, duties do not depend in this way on constitutive events; they can also exist ‘in general’. An example would be the duty not to lie. In general people have this duty not to lie, but they are not under an obligation in this respect.
Characteristic for a duty is that people can claim that the person under the duty complies with it. This claim need not go so far that compliance with the duty is enforced - although sometimes it does go so far - but it implies that people can make demands upon each other and reproach persons, including oneself, who do not fulfil their duties. Darwall writes in this connection about ‘the second-person standpoint’, the perspective people take up when they make and acknowledge claims on one another’s conduct and will. This second-person standpoint is characteristic for duties, even – be it less characteristically - for duties towards oneself, but it does not hold for all cases where a person ought to do something. For instance, if somebody is doing arithmetic and adds three and five, she ought to arrive at the outcome eight because that is the right answer. But she is not under a duty to do so. The ought at stake here is the rational ought, and normally nobody can make a claim upon somebody to act rationally for the single reason that this would be rational.
Duties are standards that determine what ought to be done. Therefore it is possible to argue from the premise that James has the duty not to lie to the conclusion that (pro tanto) James ought not to lie. Since not all oughts are based on duties, the argument in the opposite direction is fallacious: from the fact that somebody ought to do something (e.g. arrive at the outcome ‘eight’) it does not follow that this person has a duty to do so.
With the help of these distinctions, it is useful to have another look at the first analysis of Searle’s argument. On this analysis, the argument started from what was taken to express a general obligation, namely the obligation to do what one promised or contracted to do. Given that obligations are the outcome of events, such a general obligation would be a weird thing. Events are concrete, and so are the obligations resulting from them. There can be general duties, such as the duty not to steal, but there cannot be general obligations.
The formulation of the first premise of the argument
I. For all x it holds that if x has contracted/promised to do A, then x is under an obligation to do A.
can be read in two ways. It can be read, as proposed in section 4, as expressing a general obligation which can be instantiated by a concrete promise or contract. This reading is implausible, however, because it assumes general obligations.
The second reading is that this premise expresses a temporal relation between a promise or contract, and the existence of a new obligation. This is the reading which, in a more abstract form, is also used for premise VI of the second analysis, and on this reading premise I is true. But on this reading it does not express a general obligation, but it expresses the mechanism by means of which promises and contracts lead to new obligations. This is the same mechanism that is also expressed, more generally – by premise VI.
The objection that a legal ought is not a real ought might be interpreted as claiming that a legal ought is only a pro tanto or a prima facie ought and therefore not a real ought. Should not we ‘really’ do what we ought to do all things considered?
Before trying to answer this question, we must be clear about what is meant by a pro tanto or a prima facie ought? The distinction between prima facie, pro tanto and all-things-considered judgements is not always made made in a precise way and therefore it will be necessary to make some stipulations to gain the desired precision.
A prima facie judgement will be taken to be a judgement based on a limited amount of information. An example would be that somebody read the printed timetable of a train connection and arrives at the prima facie judgements that the next train will be at ten and that he ought to leave for the station in five minutes. Had he listened to the news, he would have known that there was a railway accident and that today there will be no trains anymore. The all-things-considered judgement (here in the sense of ‘opposed to prima facie’) should therefore be that the next train will not be today and that there is no need to go to the station.
A prima facie ought-judgement is an ought-judgement made under uncertainty. It is not certain that such a judgement expresses what really ought to be the case or ought to be done. However, this uncertainty is an uncertainty concerning the truth of the judgement; the judgement might be true. Therefore it cannot be said that a prima facie ought-judgement does not express a real ought. An ought-judgement that might be false is not for that reason an unreal ought-judgement.
A pro tanto judgement is a judgement on the basis of a limited set of standards. An example would be that somebody arrives at the pro tanto judgement that she ought to attend class because she was accepted for the course. From a wider perspective she ought not attend because she must go to the doctor for an urgent consultation. Since all ought-judgements are judgements based on a set of constraints, all ought-judgements are pro tanto ought-judgements. There is no ‘all things considered ought’ in the sense of an ought based on all standards, because then incompatible sets of standards should be taken into account. These would include, for instance, not only legal and moral standards, but also different legal systems with incompatible contents. If we take all sets of standards into account, we use more standards to determine what we ought to do than only the standards we are committed to. That is an impossible task, which has no outcome concerning the issue what we ought to do.
But is not it possible then to answer the question what should be done on the basis of all reasons, and is not this what ‘really’ ought to be done? This question presupposes that not all the guiding reasons identified by all the possible standards are taken into account. A more restricted set of reasons is presupposed, and presumably these would be the reasons to which the deliberating persons has committed him- or herself. The question is now a practical one, aimed at determining what to do, and not anymore what ought to be done.
Before continuing the argument, a difference in meaning will be stipulated between the words ‘ought’ and ‘should’. This difference is not reflected in ordinary usage where the two words are often used interchangeably. The word ‘ought’ will, in agreement with the analysis above, be used for what is fitting in the light of a finite set of constraints. This ought is by stipulative definition related to a set of constraints, and it does not imply commitment to the constraints, and is not necessarily related to action.
The word ‘should’ will be used for what is fitting given all the standards to which somebody is committed. Because it is based on all the standards to which somebody is committed it is in this sense used for an all-things-considered judgement. Moreover, a should-do judgement indicates that the author of such a should-judgement is committed to the standards on which the judgement is based. Normally this means also that the author is motivated to act in accordance with the should-judgement if it applies to himself, and expects compliance of other persons to which the should-judgement is applied. It is such a should-judgement which goes strongly together with the commitment to act, a commitment that is sometimes ascribed to ought-judgements.
The question that still needs to be answered is whether a legal ought can be a real ought-judgement, since it needs not express what an actor should do. Both the legal ought and the ‘should-ought’ are oughts relative to a set of standards, namely the standards of the law and the standards to which an actor is committed. Most probably these sets are not identical and this might mean that one should not do what one ought to do from the legal point of view. This has no implications, however, for the issue whether a legal ought is a real ought. A legal ought is a real ought, but most probably not a should-ought. The objection that a legal ought is not a real ought, because a real ought indicates what one should do, is based on a confusion of should-do and ought-to-do.
Given the distinction between ought and should and the observations that what ought to be the case or ought to be done does not necessarily coincide with what should be the case and what should be done, and that commitment to standards is therefore not essential for the existence of an ought, it is not really necessary to pay any further attention to one’s commitment to standards. And yet there is occasion to do so nevertheless. This occasion has to do with a link that might be made between ought-judgements and the faculty of reason. Following a Kantian style of reasoning, one might argue that not all standards count in determining what ought to be done or to be the case, but only standards that one ought rationally be committed to. For instance, some standards may be adopted on the basis of information which is recognisably wrong. Such standards have no influence on what ‘really’ ought to be the case, or to be done.
The first thing to notice in this connection is that the set of standards which ought rationally to be accepted is not by definition the set of standards which can be derived from reason alone. It is highly doubtful whether any standard follows from reason alone, but even if one or more standards would follow, it is not at all certain that these are exclusively the standards which ought rationally to be accepted. The Kantian idea that a set of standards should be derived from reason alone is a typical case of foundationalism, where reason is taken to form the foundation of all standards, a foundation which is itself above all criticism. An alternative, coherentist, view would be that all standards are subject to criticism in the light of all our beliefs and other standards, and that those standards ought rationally to be accepted which are part of a set which satisfies its own standards of rationality. This means that even if one were to embrace the idea that what ought to be done only depends on standards which ought rationally to be adopted, one is not committed to the Kantian view that reason is the unassailable foundation for these standards.
The second, and in this connection more important, thing to notice is that the rational ought is an ought relative to a set of standards; the standards of rationality this time. This ought is the ought of reason, and it fits neatly in the picture of the ought as what is required by some set of standards. There is no reason to assume that the rational ‘ought’ is the only ‘real’ ought and the fact that the legal ought needs not coincide with the rational ought is no indication at all that the legal ought is not a real ought.
Searle’s derivation of ought from rests on two steps. One step is the logical rendering of constitution. Given a setting of rules, some events lead to the coming into existence of new facts, and these facts may include the existence of obligations. This mechanism is probably the main step in Searle’s argument, and the interesting thing about it is that it has almost nothing to do with oughts or obligations. It is a mechanism by means of which all kinds of facts in the institutionalised part of social reality are generated. That some of these facts involve the existence of obligations is almost a coincidence.
The second step is the ‘derivation’ of an ought from the existence of an obligation. Like all oughts, this is a pro tanto ought. It indicates what is the only option, given a set of constraints. An obligation forms a constraint. It is a constraint on future behaviour, and as a consequence, the ought that stems from it is an ought-to-do. The possibility to derive ought from is hangs on the possibility to render this step, from the existence of an obligation to the existence of an ought based on this obligation as the only constraint that was taken into account, as a valid argument without an ought-premise. This is an issue that deserves more attention than it will be given here. Suffice it to observe that on traditional accounts of deductive logic, this is not possible. It is not possible for a very simple reason, name that in a deductively valid argument all predicators and terms in the conclusion must also be present in the premises. So if there is to be an ought-to-do in the conclusion, there must for this simple reason be an ought-to-do in the premises, for otherwise the argument cannot be deductively valid. The needed premise is the one stated above:
IV. For all x and all A it holds that if x is under an obligation to do A, then pro tanto x ought to do A.
Searle did not think this premise was very important. In his paper, he wrote that it is a tautology that, other things being equal, ‘one ought to do what one is under an obligation to do’. Some commentators have written that it is not a tautology, but rather an analytic judgement. The use of the words ’tautology’ and ‘analytic’ here suggest that something unimportant is the case, presumably based on the meanings of the words ‘obligation’ and ‘ought’. That would be an underestimation of what is involved, namely the relation between normative constraints and the impact these normative constraints have on reality in the form of what is the only proper outcome.
This relation and the way it should be rendered in a logical language are interesting topics for further study. Given the present state of logical techniques, it is probable that the introduction of some form of ought-premise will be necessary. If that is the case, it is not possible (yet) to derive an ought conclusion from is-premises. In the opinion of the present author, this impossibility has nothing to do with a gap between is and ought, but rather with underdeveloped logical tools.
This paper started with a demonstration how Searle’s derivation of ought from is could be improved. The basic idea behind this derivation, namely that certain events constitute a new ought, is correct. However, the example by means of promises turned out to be less than happy, because the argument presupposed the premise that promises ought to be kept, or something similar, and this is an ought-premise, even though it seems to be more or less analytic.
The improvement to Searle’s argument was to replace promises by legal transactions, in particular contracts. Where promises are strongly connected to obligations, legal transactions have the advantage (at least for expository purposes) that they can be used for many different kinds of changes in the world of the law, and that the creation of obligations is only one of them. The mechanism by which legal transactions bring about changes in the world of the law can be described without mentioning obligations at all, and this very mechanism can also be used to create obligations and therewith oughts. This makes legal transactions a useful tool to demonstrate how obligations and oughts can be constituted without the use of a pre-existing obligation or ought which can function as a premise of the derivation. Even though this is an improvement in the exposition of what Searle wanted to say, it still does not amount to a derivation of ought from is, because the argument still presupposes the premise that an obligation pro tanto involves an ought.
The second part of the paper deals with the objection that the derived ought is only a legal ought and not a full-blown ‘real’ ought. To address this objection it was necessary to go into some detail concerning the nature of the ought. The main conclusion in this connection was that an ought is always relative to a set of standards, and indicates what is the only allowed possibility in the light of these standards. More in particular it is not required that a real ought is what one should do all things considered, or that it should only be based on standards which one rationally ought to accept. A legal ought is what is required by the standards of the law; it does not require actual, psychological commitment to these standards, nor does it require that the legal standards ought rationally to be accepted. A legal ought is in these respects just like any other ought, and therefore a legal ought is as real an ought as can be. Therefore the derivation of a legal ought would be an adequate example to illustrate the possibility to derive ought from is, were it not that the derivation still requires the premise that an obligation pro tanto involves an ought. This premise expresses the effect of normative constraints on reality, and has as such nothing to do with what a person should do or refrain from doing. That such a premise is still necessary to model the bringing into existence of a legal ought by means of a contract has more to do with the present state of affairs in logic, than with a gap between is and ought.
At the end of this paper I want to mention briefly an issue which was not discussed yet and which might be crucial for some readers of this paper. They might be interested in the possibility to derive ought from is, because they look for an ‘objective’ foundation for duties, or for morality. This issue is then whether the derivation of ought from is shows that there are ‘objective’ duties. In my opinion, duties exist in the ‘real’ world and the derivations discussed in this paper illustrate that this is the case. However, that does not mean that duties are independent of human culture or decision making. Social reality and in particular the world of the law are both part of the real world and dependent on human culture. Therefore, in my opinion there can be ‘objective’ oughts, duties and obligations, but these objective oughts etc. would nevertheless be mind-dependent.
 The author thanks Rudolf Rijgersberg and the participants to seminars in Cracow in May 18th and in Uppsala in June 2010 for their useful comments on earlier versions of this paper.
 An example of almost critiqueless acceptance of ‘Hume’s law’ is I. McLeod, Legal Theory, 4th ed. Houndmills: Palgrave MacMillan 2007, p. 106.
 D. Hume, A Treatise of Human Nature, many editions, book III, part I, close of section I.
 P.W. Taylor, Normative Discourse, Englewood Cliffs: Prentice Hall 1961, p. 246.
 J.R. Searle, ‘How to derive ‘ought’ from ‘is’, Philosophical Review 73 (1964), p. 43-58. Also included in Ph. Foot, Theories of Ethics, Oxford: University Press 1967, p. 101-114.
 The difference between exceptions and cancelling conditions is discussed in J.C. Hage, Reasoning with Rules, Dordrecht: Kluwer 1997, p. 15-16.
 This argument against Searle’s derivation is closely related to Hare’s argument against this derivation. Cf. R.M. Hare, ‘The promising game’, Revue Internationale de Philosophie 70 (1964), p. 398-412. Also included in Ph. Foot, Theories of Ethics, Oxford: University Press 1967, p. 115-127.
 Searle did not emphasize in his paper that the created obligation is new, but from his discussion of constitutive rules in this connection, it is obvious that this is what he had in mind.
 This is argued more extensively in Jaap Hage, De wondere wereld van het recht, inaugural address University of Maastricht 1987 and in Jaap Hage, ‘What is a Legal Transaction?’, in Maksymilian Del Mar and Zenon Bankowski (eds.), Law as Institutional Normative Order, Farnham: Ashgate 2009, p. 103-121.
 Cf. R. Tuomela, The Philosophy of Social Practices, Cambridge: University Press 2002, chapter 5.
 J.R. Searle, ‘A Taxonomy of Illocutionary Acts’ in J.R. Searle, Expression and Meaning, Cambridge: University Press 1979, p 1-29.
 The example stems from G.E.M. Anscombe, Intention, Oxford: Basil Blackwell 1957, p. 56. Its interpretation in terms of directions of fit, is Searle’s however.
 In the following, I deviate from the analysis given by Searle. The reasons for this deviation are given in my paper ‘What is a Norm?, in Jaap Hage, Studies in Legal Logic, Dordrecht: Springer 2005, p. 159-202.
 Ordinary usage does not distinguish sharply between orders and commands, and the distinction between the meanings of the two terms is therefore an artificial one. There is, however, a real difference between the phenomena denoted by the terms.
 Cf. Hart’s discussion of the reconstruction of power-conferring rules as duty-imposing rules, in H.L.A. Hart, The Concept of Law, 2nd ed., Oxford: Clarendon Press 1997, p. 38-42.
Torben Spaak (personal conversation) has convinced me that, because it is a major function of the law to guide conduct, the relation between conferring a legal power and the imposition of duties is not completely contingent. Nevertheless it seems to me that rules that make it possible to create new facts (power conferring rules) are radically different from rules that impose (conditional) duties, even if the newly created facts indirectly involve the existence of duties.
 The principle ‘pacta sunt servanda’ was originally meant to express that a particular kind of agreement, the so-called ‘pacta nuda’ were binding too. Later it came to express that contracts ought to be honoured under all circumstances. Cf. Reinhard Zimmerman, The Law of Obligations, Oxford: University Press 1996, p. 576-577.
 The pro tanto means that the ought that is implied by the obligation is ‘only’ an ought as far as the obligation is concerned. There may be other reasons, next to the obligation, why the person under the obligation ought all-things-considered not do what (s)he is under an obligation to do.
 The use of brackets to denote propositional contents into which quantification is possible was inspired by different but related analyses in W.V.O. Quine, ‘Quantifiers and Propositional Attitudes’, Journal of Philosophy 53 (1956), also in W.V. Quine, The Ways of Paradox and Other Essays, revised edition, Cambridge: Harvard University Press 1976, p. 185-196, and in David Kaplan, ‘Quantifying in’, in D. Davidson and J. Hintikka (eds.), Words and Objections, Dordrecht: Reidel 1969, p. 178-214. This use requires elaboration, but this elaboration will not be given here.
Background information on quantification over states of affairs can be found in my paper ‘Reason-based Logic’ in Hage, Studies in Legal Logic, p. 69-100, in particular p. 72-76.
Those who have objections against the less than traditional use of logical techniques may find consolation in the fact that the logical analysis is not essential to the main argument of the paper.
 In a sense, because propositional content is not a linguistic expression, although it has some things in common with declarative sentences.
 Cf. the lemma ‘Use versus Mention’ in W.V. Quine, Quiddities, Cambridge: Harvard University Press 1987, p. 229-235.
 Cf. R.M. Hare, The Language of Morals, Oxford: University Press 1952, p.164-172.
 In his paper, Searle discusses this objection too, but he interpreted it as the objection that the speech act was not performed sincerely, and refuted it adequately. In this paper different interpretations of the inverted comma’s objection will be discussed, which are more troublesome.
 This assumption underlies much work in deontic logic, especially where it is attempted to construe sentences that express what somebody ought to do as sentences that express what ought to be the case.
 Apart from the ought-to-do and the ought-to-be mentioned in the main text, one may distinguish an ‘epistemic ought’. White gives in this connection the example of a law student who knows that there ought to be a law about a particular subject. This ought expresses that the existence of the law follows from what the student knows. Cf. Alan R. White, Modal Thinking, Oxford: Basil Blackwell 1975, p. 139.
 The distinction in these terms stems, to my knowledge, from Castañeda. Cf. Hector-Neri Castañeda, ‘On the Semantics of the Ought-to-do’, in Donald Davidson and Gilbert Harman (eds.), Semantics of Natural Language, 2nd ed., Dordrecht: Reidel 1972, p. 675-694.
 Cf. White Modal Thinking, p. 140.
 The possibility that physically only one possible course of action exists, is left out of consideration. Physical constraints seem to belong to an altogether different category than the normative constraints that are relevant for what ought to be the case or ought to be done.
 Cf. Joseph Raz, The Authority of Law, Oxford: Clarendon Press 1979, p. 153-157.
A legal ought can be detached on the condition that whether something is a legal standard does not depend on one’s commitment to comply with it. Raz would assume this because of his commitment to the ‘sources thesis’. Cf. The Authority of Law, chapter 3.
 With the view that there can be detached legal ought-judgements I abandon my earlier view, exposed in J.C. Hage and A. Peczenik, ‘Legal Internalism’, in pierluigi chiassoni (ed.), the legal ought, Torino: G. Giappichelli, p. 141-170. I now believe that my earlier view rested on the confusion between ought and should as exposed later in this paper.
 The nature of a pro tanto ought (i.e. every ought) will be discussed later in this section.
 Cf. Stephen Darwall, the second-person standpoint, Cambridge: Harvard University Press 2009, p. 3. Notice that, according to Darwall’s usage, a person can take the second-person standpoint towards himself, in a kind of internal dialogue.
 An exception would be that somebody has contracted with a math teacher to study properly. Then there may be a duty to arrive at the right outcome, but this duty stems from the contract and not from the rightness of the outcome.
 It might be rightly objected that this judgement is also a prima facie judgement. The problem here is that it is not well possible to give an example of a all-things-considered judgement, because all human knowledge is fallible, as it is based on a limited amount of information. The idea of an all-things-considered judgement is a theoretical construction.
 It does not necessarily mean that there is such a discrepancy, because different sets of constraints may lead to the same outcome. This would in particular be the case if somebody is committed to the law and to some other standards – for instance moral ones – and that the outcome of this broader set is that the legal standards are decisive.
 Here I tend to side with Hume (‘Reason is, and ought only to be the slave of the passions …’; Treatise of Human Nature, section II.3.3), and against Kant (Grundlegung zur Metaphysik der Sitten, p. 412-421 of the Akademie Ausgabe).
 This is the view exposed in my paper ‘Law and Coherence’.
 E.g. Hare, ‘The promising game’.
 A more extensive discussion of this issue can be found in my Studies in Legal Logic, p. 171-192.