The End-Relational Nature of Practical Reason and Morality

1. Introduction

The end-relational theory of normative language (“the Theory”) holds that all normative judgements share an end-relational nature in that they always concern how actions or objects relate to ends.  Stephen Finlay supports the Theory with analyses of the semantics of normative language and claims that the Theory is a powerful approach which gives a unified explanation of normativity, shows normativity to have a metaphysically simple nature and provides a better explanation of the nature of practical reason and morality than rival metaethical theories. [1]

This paper develops and extends Finlay’s work by showing that the Theory explains the main features of practical reasoning and then by showing how the Theory covers morality.  It also looks at how the Theory can be justified and at some of its implications.

Section 2 considers the nature of practical reason.  Using examples, I argue that the common normative vocabulary implies that practical reason has a common end-relational semantics and a straightforward metaphysical nature.  Finlay’s reductive naturalistic analyses of the words “good” and “ought” are presented, with their interesting suggestion that normative words can be reduced to probabilities.  To see how well the Theory explains the nature of practical reason, I list the major features of practical reason, and find that all of these can be well accommodated by the Theory.

Section 3 considers whether the Theory can also be applied to morality.  I list five features that morality is sometimes seen as having – self-sacrifice, respecting rules, going beyond rules, an objectified nature and a religious base.  I argue that the first three features are end-relational and contiguous with the rest of practical reason and that much of what appears to be objectified thinking in morality can be explained as due to pragmatic features of our use of language.  However, to the extent that we think of morality as mind-independent or as a divine command, this will give morality a different nature, but from a naturalistic perspective we can take the view that such thinking is erroneous and can be avoided, so that a modern naturalistic morality can follow practical reason in being entirely end-relational.

Section 4 looks at the Theory more critically and considers some of its implications.  I set out the steps involved in Finlay’s justification of the Theory and review some of the criticisms that have been made.  I then offer a further argument for the Theory – that for any normative judgement it always seems possible to ask: “for what end?” – and suggest that this shows that end-relativity is a deep feature rather than just a matter of semantics.  I then consider the end-relational theory in comparison with other theories of metaethics and argue that it provides the best explanation for our understanding of practical reason and morality. Finally, I reflect on some implications of the Theory’s binary object-ends conception of normativity.

2. Practical Reason and Metanormativity

2.1 Defining Practical Reason 

This paper starts with practical reason outside morality as it seems more straightforward than morality.  Practical reason, choosing how to act, is a central feature of human life.  We always need to decide what to do – on matters mundane or profound, as individuals or as groups.  By contrast, morality can be placed as a relatively minor element within the broad field of practical reason.  Yet, historically the philosophical focus has been disproportionately on morality rather than on practical reason, for both normative enquiries into what morality should recommend and metaethical enquiries into the nature of morality.

In this paper, by practical reason is meant choosing how to act generally. As such it is a very wide subject covering all choice-making.  It is all acting and thinking where the direction of fit is to change the world rather than to receive information from the world and is a basic human function.

Practical reason is a part of the broader domain of normativity.  The normative covers evaluations and value judgments, how things should or ought to be and relations to evaluative standards.  Normative claims can be contrasted with positive claims which cover descriptions, explanations, predictions and facts.  Practical reason is the part of the normative that is concerned with evaluation of the practical, of actions. 

2.2 Normative Language

 In trying to understand the nature of practical reason, a starting point is to note that a simple normative vocabulary is used throughout practical reason. 

For example, the word “good” can be used for simple individual decisions such as “it would be good to have a coffee”; for government decisions “it would be good to reduce taxes”; and for morality “it is good to give to charity.”  The word is also used for normative assessments not relating to action and therefore outside practical reason, such as in functional judgements “this car is good,” in epistemic judgements “this is good evidence,” and in aesthetic judgements “this is a good painting.” [2]

The basic normative vocabulary includes ubiquitous words such as “good”, “ought” and “right”, which are related to each other and convey a suite of “thin” general purpose normative concepts that we readily apply across numerous contexts.  We also have words that are “thick” in that they combine normative and factual elements, for example “quick”, “kind”, “brave” and “stubborn”.

It is striking that our normative vocabulary is used easily and confidently by speakers as good tools employed instinctively without conscious awareness.  This motivates the approach of this paper of trying to understand the solidity of our use of language in practical reason and see if this can be extended to morality.  

2.3 End-Relational Theory

What is the common meaning of such widely used normative words? The end-relational theory of normative language proposed by Finlay suggests that they are all about making an evaluation of an object, state or action relative to a norm, end, standard, interest, or other criteria.[3] The Theory claims that we use the words fluently as we instinctively apply an underlying end-relational semantics but that the underlying semantics is often obscured as we communicate pragmatically and do not state explicitly elements that are clear from context.  For example, I say “I ought to go home now.”  The Theory suggests that this is a statement relative to an end which is omitted by ellipsis such as “so I can get to bed.”  The end or norm may be singular or multiple and may be intermediate or ultimate; it is commonly vague and fluid.[4]  But the end-relational theory suggests that the underlying semantic of the normative words is always about evaluation of an “object” (broadly construed to include actions and states of affairs) relative to an “end” (broadly construed to include norms, standards, interests and other criteria).[5]  So, the core meaning of “good” (the most general normative adjective and adverb) is that an “object” increases the probability of an “end”, and the core normative meaning of “ought” (the most general normative verb) is that an action or other “object” is the most probable route to achieve an “end”.

To illustrate: my end is to clean a very dirty pan which I know from experience would make a mess in the dishwasher.  I naturally employ the normative vocabulary – I “ought” or “should” or “must” wash it by hand, washing up would be “good” or “better” or “best” or “right” or “reasonable”;  I “ought not” or “should not” or “must not” use the dishwasher, to do so would be “bad” or “worse” or “worst” or “wrong” or “unreasonable”.  My normative judgement that washing up is the better route to the end of cleaning the pan is a straightforward answer to a question about the relation between an “object” (using the dishwasher) and an “end” (cleaning the pan) which, having experienced a bad pan messing up a dishwasher, has a clear factual answer.

The picture of practical reason is of action being guided by binary evaluations of things relative to criteria. Trivially, any binary evaluation could be made – “Are bananas good for high tides?” answer: “neither good nor bad” – but we concentrate on evaluations that we consider relevant.  Evaluations are never made in isolation but always relative to something.

Finlay’s phrase “end-relational” is misleading as it implies a narrow sense of “end” as, broadly, “desired state of affairs”.  I would prefer to say “criterion-relational” to clearly show a wider sense of “ends” as, broadly, “principle used in judging.”  Looking at dictionary definitions, it is “criterion” rather than “end” that means anything against which evaluations are made. Misreading “end-relational” as concerning the narrow sense of end may imply that the Theory is consequentalist, but this is not the case: the Theory covers all normative evaluations including assessing actions relative to rule-based or virtue-based standards.    For example, in rule-based ethics, actions can be assessed relative to rules, and rules themselves can be assessed relative to criteria such as coherence with our intuitions; both of these evaluations are “criterion-relational” although not narrowly “end-relational”.

The relational nature of practical reason is not obvious partly because assumed “ends” are omitted by ellipses but also because “ends” are commonly obscure, vague, fluid and multiple.[6]  It is efficient not to specify “ends” more than necessary:  in choosing what to eat I may have multiple objectives including health, cost and taste, but will not normally analyse closely how these may be balanced.  Often my “ends” are not specified further than “good for me generally.”

It is commonly argued, following suggestions in Hume[7], that it is not possible to reason about ends, but the obvious fact that we do debate about ends (in both the wide sense  of “criterion” and the narrow sense of “desirable state”) shows this to be false.[8]  To the contrary, deliberation about “ends” is as much part of practical reasoning as is deliberation about means. As the Theory holds that all normativity takes an end-relational form, so all evaluation of “ends” will itself be in end-relational terms. Any “end”, even an “end” that someone considers to be an ultimate end, can be evaluated relative to an infinity of other criteria.  

We sometimes say: “good all things considered.” This implies that all relevant means and ends have been considered and appropriately weighed.  However, apart from in simple cases, consensus should not be expected as there will be different views on what “ends” and means are relevant and how they should be evaluated and weighed.

2.4 Analysing Normative Words

As analysed by Finlay, under the end-relational theory “good” is a binary predicate with two principal terms, an “object” p and an “end” e.  If we say “it is good to process a bad pan by hand” this is elliptical for “it is good to process a bad pan by hand to clean it” with a logical form of Good(p,e).  “Good” deals with the relationship between p and e, approximately that p promotes e.  The relationship is assessed assuming a factual background b, which here would include assumptions about the normal capabilities of dishwashers.

Finlay thinks we can go further and define “good” as meaning that p increases the probability of e, so: 

It’s good for e if p given b means that pr(e\p & b) > pr(e\~p & b)

p is good for e for given background b if the probability of e occurring with p is greater than the probability without p.[9]  Or “cleaning a bad pan is normally more likely by hand.”

While “good” is the most general normative adjective, “ought” is the most general normative verb.  “Ought” is one of a series of modal verbs which indicate variants of necessity (“must”, “have to”) and possibility (“may”, “can”).  Necessity modals mean roughly “in all possibilities” and possibility modals “in some possibilities”.  “Ought” and its near-synonym “should” generally refer to the most likely.  

Modal verbs have both normative and non-normative uses and Finlay considers that we recognise normativity when there is an end to be addressed.  This can be shown by how we use the tenses of the modal verb “have to” to indicate normative readings.  The sentence:

            You have to use the key to open the door

is normative, recommending an action to be done, but

            You had to have used the key to open the door

is epistemic, reporting on knowledge of an action taken.  The tenses of “have to” are employed in the first sentence to show that the use of the key will be instrumental and in the second sentence that the key is no longer instrumental. 

Finlay proposes a probabilistic semantics of normative uses of ought:

            One ought to p for e given b means pr(e\p & b) > pr(e\r & b) for all r in R

One ought to take action p for the end e given background b means that the probability of e is higher with pthan for any alternative r in the set R of relevant alternatives. Or, the meaning of ought is to take the most likely means to one’s end.[10]  This also equates to “take the best means to one’s end” because “best” is just the superlative of “good” and “good” is a probable means to an end.  “You ought to process a bad pan by hand to clean it” means “processing a bad pan by hand is normally the most probable (=best) way to clean it.”

There is an interesting result here, that for at least some examples of practical reason a reductive analysis of normative words in terms of probabilities appears possible.  Finlay believes he has produced a naturalistic analytic reduction of normative words, contrary to Moore’s view[11] that such a reduction is impossible and to “Hume’s Law”[12], the doctrine that the circle of normative words such as “ought” cannot be derived from factual words.  (Chrisman 2014) notes that it is questionable whether subjective probability is a scientific rather than normative or expressive concept, but, apart from that, the reduction to probabilities looks promising.  However, the challenge from both Hume and Moore is about moral rather than instrumental uses of “ought”, which I will consider later.

2.5 Strengths of the End-Relational Theory

Returning to the dishwasher example, my normative judgement that the dishwasher is not the best route is a judgement about the relation between an aspect of the world and an end.  This is similar to a factual judgement such as “there is a dishwasher in the kitchen” which relates something in the world to human concepts.  Another example of the closeness of positive and normative judgements is the similarity between a positive prediction that the sun will rise at 6am tomorrow and a normative judgement to travel after 6am to drive in daylight.

These examples show that the end-relational theory is powerful not only because it provides a unified explanation for the wide range of uses of normative words but also because it can be seen as presenting normative statements as having a similar nature to factual statements.  Both factual and normative statements consider objective features of the world using our concepts.  A factual statement such as “there are two cakes on the table” uses our concepts to report on aspects of the world.  A normative statement like “eating both cakes would not be good for the health of a diabetic” uses our concepts to evaluate how “objects” in the world relate to “ends” and thereby provides normative guidance about how to act.  As factual judgements vary from the certain to the unknowable, so normative evaluations can range from the obviously true “I must drink to stay alive” to the unknowable “I don’t know whether taking this job will be good for my long-term happiness.”

On the end-relational theory, the metaphysics of normative language is therefore simple and uninteresting.  A normative judgement is just a recognition of a relationship between an “object” and an “end”.  It is no more interesting than a recognition of a fact in the world.  A value judgement is not a “thing” that exists in the world beyond the judgement itself, it has no more of an ontology than does the recognition of a fact. This can be considered a kind of realist view in the sense that normative judgements are considered as real as factual judgments.[13]  Although both involve human assessment, they both are tied to the world.

Importantly, the end-relational theory seems to cover the key features we see in normative language and thinking:

1. Normative judgements are truth-apt or cognitive – judgements about “objects” relative to “ends” can be true
2. Normative judgements have a simple, non-mysterious metaphysics – just recognition of a relationship between “object” and “end”.
3. Normative judgements have a simple epistemology – we know normative truths by making evaluations of how “objects” relate to “ends”
4. Normative judgements are made by all – as deciding how to act is a basic need
5. Normative judgements range from certainly true to totally debatable – depending on the normative question asked; some answers converge, others can’t
6. Normative judgements are in a sense objective – as they relate to facts in the world and use reason
7. Normative judgements are in a sense subjective – as they answer human-made questions and commonly relate to human ends
8. Normative judgements are generally motivating – as they generally relate to chosen ends that matter
9. Normative judgements use reason – as the reasons for and against a judgement are assessed.
10. Normative judgements supervene on facts in the world – as judgements depend on how facts relate to “ends”.[14]

It is impressive that the Theory can comfortably explain all these features, unlike rival approaches.  I will look at the Theory more critically later, but first I will consider whether it can also be applied to morality.

3. Morality and Metamorals

3.1 Placing Morality

Morality has proved difficult to define.  For example, it has been distinguished as being about duty[15], or obligations[16], or universal prescriptions[17], or universal rules[18] or categorical imperatives[19] or reactive attitudes[20].  Although some writers have seen one or another single feature as definitive of morality, highlighting a single aspect is unlikely to do justice to our conceptions of morality, which may be complex, varying and vague.  Instead, I will try to place morality within practical reason by considering five major features. 

The first feature is that morality is about overriding one’s own immediate interests for the sake either of others or of some principles.  So, morality requires me to forbear from stealing even when it would be in my interest, either from concern for the owner or from some sense of principle, and morality also requires me to help when there are emergencies such as lives at risk.   While the “self-sacrificing” feature of morality is mainly about the needs of others, for some morality can also require putting principles such as sexual purity and religious observation ahead of personal desires.  That morality provides motivation to override personal interests can make it seem important, precious, other-worldly and mysterious and may contribute to the tendency to see morality as being more than just an aspect of the natural world.

The second feature is that morality concerns a range of central rules.  Most societies seem to have in some form a suite of basic rules and restrictions which mainly relate to human welfare, which from evolution come naturally to us and which are widely accepted.  Compliance with such “core morality” is supported by “reactive attitudes”[21] – failure to comply leads to attitudes such as blame, resentment, punishment and guilt.  “Core morality” appears essential to the functioning of society by constraining damaging behaviour, removing fears of violations, promoting a degree of altruism, supporting agreements and property rights, upholding family obligations, and providing coordination.

But morality is sometimes seen as being about more than just following core morality.  The third feature which can be termed “aspirational morality” relates to morality’s wider concerns including altruism beyond the minimum required by core morality, concern for non-human animals, religious considerations such as religious worship and practice and personal principles such as keeping a healthy body and mind.  While core morality is about near-compulsory basic behaviours needed to make society function, aspirational morality includes other principles that are important to individuals, but which are less supported by the reactive attitudes and not compulsory in the same way.  Aspirational morality can include thinking beyond one’s self and immediate circle towards considering what would be for the best “all things considered.”

While the first three features concern the function and content of morality, the fourth feature concerns some distinctly objective and absolute concepts and language which are often used in morality.  I will distinguish four related points:

• Categorical imperatives. Sometimes we speak of moral behaviour as being required absolutely, not relative to any ends:  we say “do not steal” as a categorical imperative rather than, say, “do not steal if you want to avoid suffering” as a hypothetical imperative.
• Law-like character. Morality has law-like features with rules that one is under an obligation to follow.
• Objectified language. Morality is commonly thought of as an important objective requirement that deserves respect and from which one should not lightly retreat. Sometimes our language suggests that morality is about independently-existing objective facts, as when two sides in a moral debate will claim that they are each “right”, implying there is an objective fact to be found.
• Mind-independent existence. Sometimes moral requirements are believed to have some kind of metaphysical existence independent of ourselves.

The final feature is morality’s relation to religion. Historically and currently, many see morality as commanded by a god, and some may also think that obedience to morality is supported by rewards and punishments in an afterlife.  For many believers, morality will have a different metaphysics from the rest of practical reason – a command from god rather than a part of   deliberation about how to act.  In addition, religions often extend the content of morality beyond welfare to religious worship and practice, sexual behaviour and diet.  Even for the non-religious, elements of the divine-command thinking may still remain in some moral language and concepts, such as seeing morality as a binary good versus evil matter and thinking of morality as something mysterious and unworldly.

How widespread are these features?  Overriding one’s immediate interest at times seems to be an almost universal evolved feature, as does the presence of central rules and a degree of wider concerns.  By contrast, although they are common, categorical imperatives, law-like character, objectified language, belief in independent existence and a relation to religion are all features that vary between cultures and individuals.

So, in summary, morality is a part of practical reasoning which concerns putting principles or other’s needs above one’s immediate desires by complying with a set of core rules and sometimes going further.  It can use distinctly objectified language and is often associated with religion.

3.2 Applying the End-Relational Theory to Morality

Having identified five features as distinctive of morality, we will now consider whether each of these can be accommodated by the end-relational theory.  The first three features all concern the content of morality as making recommendations and rules that affect action. Looking at these together, in respect of its action-affecting content, morality can be seen as like other practical thinking but with a particular focus on requirements to override self, to consider others, to follow principles and to comply with rules.  All of these elements can be dealt with under the end-relational approach as the needs of others, principles and rules can be included among the ends to be pursued.  Also, these elements are not unique to morality – for example we need to consider others in our families, and we need to comply with rules in our jobs.  As in my marriage I can think “I ought to cook the meal for the end of helping my partner,” within morality I can think “I ought to hand in a lost wallet for the end of helping the owner.” Similarly, as at work I can think “I ought to protect confidential information for the end of following the rules of my job” so within morality I can think “I ought not to steal for the end of complying with morality.” So far as morality is about self-sacrifice, considering others, following principles and respecting rules, this can be accommodated by end-relational thinking, as occurs elsewhere in practical reasoning outside morality.  As far as morality’s action-related content is concerned, there seems to be considerable continuity between morality and practical reason generally. Deliberation about morality can be conducted in the same end-relational terms used elsewhere using the common normative vocabulary.

But while morality’s content can be dealt with in end-relational terms, the fourth feature, its objectified language, looks much more challenging.  Historically, most writers have seen morality’s objectified and categorical language as the distinct mark of the moral and something which definitely cannot be assimilated to an end-relational model.[22]  However, drawing on Finlay, I believe that the apparently objective elements can be accommodated.  Looking at the four objectified features in turn:

• Categorical imperatives. Finlay explains in detail that what appear to be categorical moral statements actually reflect ends omitted by ellipsis.[23] He explains that the ends are sometimes mutually agreed between conversational participants and sometimes imposed rhetorically by the speaker. Finlay’s analysis suggests that despite initial appearances, categorical imperatives are actually expressed relative to ends that are not explicitly stated because they are understood.
• Law-like character. Morality’s law-like nature can be explained as part of its core morality function. As an important aspect of morality is following core rules, it is not surprising that morality often has a law-like appearance, with a strong presumption backed by reactive attitudes that rules of core-morality should be followed. But this does not necessarily imply that these rules have a special nature, such as having a mind-independent existence.  A law-like character on its own is not incompatible with morality having an end-relational nature. 
• Objectified language. Someone says that “eating meat is morally wrong” and his colleague disagrees. The language seems to suggest that there is a fact about what is morally wrong somehow “out there” to be discovered.  But this is not necessarily the case.  The language may just be used for convenience and rhetorical force without the participants being committed to any view about the metaphysical nature of morality.  I can express myself more concisely and strongly by saying “eating meat is wrong” rather than “in my view eating meat causes suffering and should be avoided” but this does not necessarily mean that I am referring to an independently-existing moral rule. 

But even if the objective features are only trappings, they are an important feature that gives morality its “clout.”  Committed participants in moral disputes may regard their views as objective requirements and so may simply stick with non-relativized views such as “eating meat is morally wrong.”  But it seems to me that there is always the option of saying more about the end-relational reasons that support a moral position.  Indeed, discussions about moral issues will often move to end-relational terms as it is natural to ask “why?” about a disputed moral requirement and to try to reason towards a consensus.  This does not mean morality becomes relative or a matter of taste – we can still be strongly committed to a moral view, but rather than relying on a bald categorical statement we have the option of using end-relational reasoning to provide support for our position or to engage with the views of others.   Even where language appears objectified, as we can ask “why?” end-relational thinking is not far below the surface.

• Mind-independent. But sometimes morality is indeed thought of as having a mind-independent existence.  It should be acknowledged that within this perspective, morality may not be end-relational but instead be regarded as an independent requirement.  However, to regard morality as having an objective, motivating, mind-independent existence seems incompatible with a naturalistic outlook and so, for a naturalist, may be regarded as an error. [24]

Turning finally to the fifth feature of morality, its connection with religion, it should be recognised that some religious conceptions of morality are not end-relational. On a divine command view, morality has a non-natural basis and has a different metaphysics as a command from god rather than relating to human recognition of relationships between facts and ends. But, again, divine command seems incompatible with a naturalistic perspective and so from this viewpoint such thinking can be regarded as an error.

3.3 Do Practical Reason and Morality have the same Nature?

Even without considering end-relativity, I see three main reasons to expect that morality will have a similar nature to other parts of practical reason.  The first is that the underlying subject matter is the same – both are ultimately concerned with how to act. The second is that they both share the same normative language.  The third is that there is continuity between morality and practical reason with no agreed way to draw a boundary.

The end-relational approach strengthens the case by revealing the nature of practical reason and suggesting that morality may have the same nature.  If the semantics of our normative language is normally end-relational, then it would be expected that in dealing with matters in the moral domain we will often work in end-relational terms.  This is indeed the case, as we can and do deal with matters of self-sacrifice, core morality and aspirational morality in end-relational ways as when we say: “you must not steal because of the harm it will cause.”  It seems that even with moral matters, our default ways of thinking are end-relational and that we always have the option of thinking in end-relational rather than in absolutist moral terms.

Yet it is very commonly thought that morality is different, perhaps because of its objectified and absolutist features, its relation to religion and its ability to motivate self-sacrifice. My response to this has three elements.  First, I would point out that morality is similar to practical reasoning and it often uses end-relational thinking.  Second, I would argue that the apparently absolutist features of categorical imperatives, law-like thinking and objectified language can be explained as pragmatic features that do not necessarily imply commitments to a mind-independent conception of morality.  And third, I would argue that to the extent that morality is thought of as having a mind-independent objective existence or as being based on divine command, that this is an error and is a way of thinking that should be avoided. Therefore, on my approach, morality can be seen as end-relational, apart from certain elements of thinking which can be regarded as mistaken.

3.4 Moral Objectification as Optional

We are able to choose not to think of morality as being based on divine command or as being mind-independent because our moral reasons are overdetermined.  It is sometimes claimed that if we stop thinking of morality in religious or objectified ways then it will be undermined, but this is not the case as we have a number of other ways of dealing with moral matters.  Consider the possible motivations to act morally, say where I could take cash from someone’s wallet without being found out. I may act morally from:

• Instincts and socialisation – without even thinking about it, taking the money is something that I just don’t do.
• Reactive attitudes – I fear being found out.
• Empathy – I have sympathy for the other person and won’t violate their property.
• Universal consequentialism – I reckon the other person’s loss will exceed my gain.
• Virtues – I don’t want to be the sort of person who steals.
• Constructivism – I recognise the “do not steal” rule is advantageous overall.
• Mind-independence – I consider the “do not steal” rule to be a mind-independent requirement.
• Divine command – I do not steal because this is commanded by God and I either respect God or fear hell fire.[25]

The first six considerations remain even if we stop thinking of morality in objectified or religious terms.    Our general normative vocabulary with words such as “good” and “ought” can provide a sound resource for dealing with moral matters in a non-objectified and non-religious style.[26]

If some religious or mind-independent conceptions of morality are wrong, and if categorical imperatives, law-like character and objectified language only have a pragmatic function, then there may be scope to revise our moral practices. We can consider which of our absolutist moral approaches are unhelpful and should be avoided and which should be retained in some form.  Here I would distinguish between moral thinking and moral rules.  With moral thinking, using objectified moral concepts gives morality clout, but can confuse our thinking and so should be used with care.[27]  By contrast, with moral rules there are advantages for motivation and coordination to having clear rules and therefore there is a case for retaining moral rules as a useful fiction.[28]  The function of morality isn’t optional, the way to think about it is.

It is useful to consider a practice similar to morality which operates without objectified concepts.  Morality borders what may be called “civility”, being reasonably polite, respectful and thoughtful in day to day interactions.  I take my place in a queue, I respect personal space on the underground, I am patient listening to a work colleague’s stories and so on.  Civility is similar to core morality as a valuable evolved social feature supported by reactive attitudes.  But when listening dutifully to a colleagues’ stories I wouldn’t think that there is a mind-independent or God-given standard of polite patience to be followed. Civility operates without invoking mind-independent objectified rules or religious commands and we can and do deal with morality in the same way.  

Conceptions of morality change over time.  The Greeks did not have our idea of the moral[29]  and current conceptions have roots in particular religious and philosophical ideas.    It may be that our moral concepts are in the process of evolving so that they become more thoroughly secular and make morality less distinct from the rest of practical reason.    If practical reason has an end-relational nature, then a modern morality can too.

4. Review of the End-Relational Theory 

4.1 Evaluating the End-Relational Theory

Having first considered the strengths of the End-Relational Theory as an explanation of practical reason, and then its application to morality, I will now look further and more critically at the Theory and consider some of its implications.

I will start by evaluating how Finlay justifies the Theory.  The steps needed in his argument can be set out as follows:

1. A distinct normative vocabulary is used confidently by speakers across normativity
2. Humans have an extraordinary capacity to apply grammars and semantics[30]
3. It can be postulated that there is a simple semantic applied generally to normativity.
4. It can further be postulated that this unifying semantics is end-relational.
5. Review of linguistic evidence supports (3) and (4).
6. Evidence to the contrary can be dismissed by appeals to pragmatics, in particular
   1. A normative judgement not relativized to ends will have an implied end omitted by ellipsis.
   2. Where a normative judgement has motivating force, this is due to pragmatic factors
7. Underlying semantics rather than pragmatic factors determine the nature of normative language.
8. Language rather than non-linguistic factors determine the nature of normative thought
9. The nature of our normative thought rather than other factors determines the metaphysical nature of normativity.
10. As normativity is only about recognition of relationships between “objects” and “ends”, and assuming “ends” are purely naturalistic, normativity has a simple naturalistic metaphysic.

The first two points, the existence of the normative vocabulary and the recognition of our amazing human skills with syntaxes, launch Finlay’s project.  We can agree that it looks promising to consider the extent to which different normative uses have something in common.

On the third point, Finlay argues strongly that our confident use of a word like “good” in so many contexts can only be explained if all the uses have a single underlying meaning.  He may be correct that many uses of “good” have related meanings.    But, as pointed out by (Fogal 2016), it is rare for words to be entirely monosemous (with only one meaning) and most words are to a degree polysemous (with more than one related meaning).  This is indeed the case for “good” which the Oxford English Dictionary shows as having around a dozen principle meanings, some of   which are not end-relational including “goods” as merchandise and “good grief!” as an exclamation.  Such particular uses need not undermine Finlay’s main contention, that over a wide range the word “good” has an end-relational semantics.  But the likelihood of polysemy weakens his justification for insisting that all uses should be reduced to end-relational forms.

Points 4 to 6 are the main focus of Finlay’s work.  He explains at length how the end-relational theory is supported by linguistic analysis and how pragmatics can be employed to explain features that initially may appear incompatible with the end-relational approach.  He makes the case that the end-relational semantics apply widely, and he finds linguistic arguments to bring potentially conflicting uses under the end-relational model.  But his evidence for assimilating all uses as end-relational cannot be definitive as his linguistic analyses are by their nature somewhat arbitrary and alternative explanations could be proposed. A central end-relational semantics for normative language appears to be well supported, but the evidence is inconclusive on whether all normative uses are end-relational.

Point 7 is an important part of Finlay’s methodology.  His view is that in understanding normative language we should concentrate on the semantics, which he takes as straightforwardly end-relational, and make correction for pragmatic features, particularly by reinstating ends pragmatically removed by ellipsis and by ignoring pragmatically added motivating force.  This leads to his view that normative language is fundamentally end-relational, and its motivating features are added optionally.  He therefore dismisses the motivating character of some normative communication as just a pragmatic surface feature which he labels “quasi-expressivist.”  But the expressivists would have it the other way around – for them normative language is fundamentally about expressing attitudes and the cognitive appearances are a surface feature – “quasi-realism”.  I think Finlay has it right here.  The semantic meaning of a normative phrase should be our focus as it does not vary between uses, unlike the expressive elements which vary between uses, the Frege-Geach problem. [31]  Compare a speaker stating, “grapefruit is good to eat” with the same phrase when embedded as in “some people agree that grapefruit is good to eat”.  With the first-person statement we make the pragmatic assumption that the speaker favours and is motivated to have grapefruit while in the second these pragmatic elements disappear, and we just have the underlying semantic meaning – a favouring relationship between grapefruit and eating.

Finlay says little about Point 8.  But, as argued in (Gregory 2015), it is possible that there is more to normative thought than to normative language.  For example, we may think in unrelativized terms when obvious ends can be taken as given, and categorical “oughts” may play a psychological role in closing mental deliberation.  Finlay suggests that pragmatic features extend to thought, such as when I think sarcastically “my boss is really bright” or think rhetorically “his clothes are an abomination”, but Finlay does not seriously address how normative thought may differ from normative language. It may be a reasonable assumption that if end-relational thinking is dominant in language then it will also be a major feature of thought, but it is difficult to dismiss the possibility that normative thought has additional features.

Finlay also says little about Point 9.  Why should our language or our thoughts reflect the real metaphysical nature of normativity?  Even if our language was purely end-relational this would not rule out reality being something different, for example that what should really be done should be based on supernatural requirements.  Finlay could respond to this by saying that provided a naturalistic position is assumed, then normativity will be nothing more than a human practice, and the end-relational theory provides a good account of this practice.  He presents a coherent picture of normativity as a natural practice rather than arguing directly against non-naturalistic views.

Finally, Point 10 introduces a further problem. Even if it is granted that normative language is end-relational, the issue still arises of the ends or other criteria to be used.[32] If I ask “why should I not steal” you may reply in end-relational terms with a consequentalist reason “for the end of avoiding harm” or a constructivist deontological reason “for the end of respecting a rule that is generally beneficial”.  But you could also just say: “for the end of following what is required by morality”. Because we can take “morality” or “obedience to god” as a criterion, the end-relational theory on its own cannot determine the nature of morality and cannot eliminate conceiving of morality as having an objective and perhaps non-natural existence.

What should we make of these various criticisms?  They do not undermine the view that much of normativity is end-relational, but they make it more difficult to definitively dismiss alternative conceptions. 

But I consider the end-relational theory can be further supported by another argument. This is that it always appears possible to question evaluations in practical reasoning by asking “for what end?” or “by what criterion?”  This suggests that under questioning all evaluations turn out as needing to be related to some end or other criterion.   Although we may initially think of an evaluation as unrelativized, asking the question always seems to tease out relationships to ends or other criteria.  An evaluation that under questioning stood in isolation and could not in some way be related to some end or other criterion would be a strange animal.  I find this a strong argument that cuts through the criticisms made of Finlay and suggests that the end-relational nature of normativity runs deep.

4.2 Does Practical Reason have to be End-Relational?

If we grant that normative language either sometimes or always has an underlying end-relational semantics, an important question then arises.  Is this just a formal grammar-like feature or does it reflect the way in which normative thinking has to be?

As an example of a purely grammatical feature, take the phrase “it is raining.”  English grammar has a common agent-verb-object pattern, so to say “raining happening” within this form we use “it” as a notional agent to hang what we want to say over the grammatical frame. The agent-verb-object base presumably comes from appreciating our own agency, but we can recognise that grammar does not reflect reality in a simple case like “it is raining”

Might the end-relational semantics of normativity similarly be a formality that does not represent reality?  Arguably, normative language and thinking is more likely to reflect reality because, at least on a naturalistic view, normativity is at root only about human thinking. If there is nothing beyond our thinking, then understanding our ways of thinking may be all there is to understanding normativity.

Does normative thinking have to be related to “ends”?   The strong view is that we cannot work in the normative domain without reference to some end or other criterion, whether explicit or implied.  On this view, an evaluation cut off from “ends” would be senseless in terms not only of semantics but also of what a normative evaluation fundamentally is as a measurement against a standard.

The point mentioned above is important here – our apparent ability to question any normative evaluation by asking “for what end?” or “by what criterion?”  I said an evaluation that could not be related to some end or other criterion would be a strange animal.  This strangeness seems to run deeper than just language: whatever words are used in making an evaluation, it would be odd to have a normative evaluation that could float free from any questioning about ends or other criteria.   

To challenge this view, let us review cases of what appear to be unrelativized evaluations.  I earlier considered moral categorical imperatives and concluded that despite initial appearances they are related to “ends”.     Another kind of normative judgement that appears to be absolute is a judgement about ultimate ends.  It may be thought that while instrumental ends are by their nature relative, eventually we come to ultimate ends which are not based on other ends.  As an example of a statement of an ultimate end take “happiness is good for its own sake.”  Finlay analyses this as involving a “reflexive relativiser” in the form “p is good for p” so that the statement means “happiness is good for happiness.”[33] Because under his theory “p is good for p” means “p increases the probability of p”, this is trivially true, and we can say the same about something with only instrumental value like “medicine is good for medicine” or something intrinsically bad like “pain is good for pain”.  But Finlay suggests that we say “happiness is good for its own sake” to pragmatically communicate that contrasting with instrumental uses, happiness has a positive value that is non instrumental. In support of this analysis he notes that we do not say “pain is bad for its own sake” which would mean the untrue “pain decreases the probability of pain” and instead we say “pain is intrinsically bad”. 

Finlay’s analysis is interesting, but ultimate ends can be more straightforwardly accommodated by noting that they are only ultimate on a narrow conception of end.  When we describe happiness as an ultimate end, we are using end in the narrow sense of “desired state of affairs”, we mean that happiness is a state of affairs that is valued intrinsically and not in terms of other states of affairs.  Saying that happiness is an ultimate end is the end-of the line for “desired-state-of-affairs” thinking but not for broader “criterion-relational” thinking.  We can still evaluate happiness against any criterion we choose. Trivially, we could ask if happiness is good for bananas, and more usefully we can ask questions such as how good happiness is relative to other objectives. It appears to me that substantive further evaluations can always be made of claimed ultimate ends.  For example, we can ask whether “utilitarianism is good all things considered”, “obeying God is good all things considered” and even whether “doing what is best all things considered is good for the end of my immediate pleasure”. Although there may be chains of normative evaluations of states of affairs leading through instrumental goods to ultimate goods, these do not exhaust the range of criterion-relational evaluations that can be made.  In response to “happiness is good” we can still ask “good for what?”  Views about what is intrinsically good are always controversial and opponents will expect a justification to include an analysis of what it is good for. It seems that contrary to initial appearances, what we regard as intrinsically good does not reflect unrelativized value. 

Let us consider some other normative statements which appear to be unrelativized. Is “one ought to be practically rational” absolute?  No, because we can ask ‘for what end”’ and the response is ‘for the end of achieving your “ends”’, as opposed to, for example, the end of acting spontaneously.[34]  Is “one ought to act morally” absolute?  No, it is true if the end is to follow morality and not true if the end is self-interest.  Similarly, “one ought to follow god’s commands” is true if the end is to follow god’s commands and not true if the end is immediate gratification. A confusing feature that these three normative statements share is that when relativized they are reflexive, as in “you ought to act morally to act morally”.  But this does not stop them being end-relational, as is demonstrated when questioning shows that other ends can be conceived. For each of these statements we can debate whether or not to follow the self-reflexive end assumed, showing that these statements do not stand in splendid isolation from the scope of end-relational evaluation.  

I conclude that the suggested cases of unrelativized normative thinking can on reflection be seen as end-relational and that we should accept the strong view that normative thinking has not just an end-relational semantics but also an end-relational metaphysical nature.

But practical reason and particularly morality are often thought of in absolutist rather than end-relational terms.  According to the end-relational theorist, what kind of error are the absolutists making?  Finlay[35]argues that our practical reasoning proceeds well on instinctive end-relational terms and that problems only arise when we reflect on what we are doing.  Our first-order practical decisions are fine, it is only our second-order understanding of our normative concepts and the metaphysics of normativity that go wrong.  To the extent that we think of normativity as absolute we are making a mistake, but Finlay considers that our default background thinking will remain end-relational so that our errors in self-understanding are limited and not fundamental to what we decide to do.  As in navigation, the fact that all positions are relative does not invalidate the conclusions of a navigator who think of positions as absolute, so thinking of practical reason as absolute does not stop us making effective choices. While faulty theories of the nature of practical reasoning will have some effects on behaviour, for example when someone acts in accordance with a religious code, Finlay argues against Mackie[36] that in general objectified concepts are a limited aspect of practical reasoning and morality.

4.3 Comparison with other Metaethical Views

Finlay claims that the end-relational model provides a theory of practical reason and morality which overcomes central issues that beset both primitivist moral realism and moral expressivism.[37]  He considers that the end-relational approach includes the insight that normative thinking is two-dimensional as it considers relationships between objective facts in the world and subjectively chosen ends.  By contrast, the other metaethical views take one-dimensional approaches, in that primitivist moral realism concentrates excessively on the objectivity and the cognitive content of the facts, while expressivism concentrates excessively on the subjectivity and motivating character of the ends.

Finlay also claims that the end-relational approach explains why some normative thinking matters and motivates while other normative thinking is merely formal and does not involve motivation.[38]    Both the following normative judgements may be true –  “I ought to drink to avoid dying of thirst” and “To follow the rules of chess, a bishop ought to be moved diagonally” – but the first concerns an end that matters and is of importance while the second does not. According to Finlay, what is important depends on context.[39]  Commonly what is important are the reasons relating to the subjects’ internal motivations and to an extent internalism about reasons[40] is true as internal reasons commonly motivate.  But the externalist is right that not all reasons that are important are internally motivating.  Normative judgements can consider all kinds of relationships between “ends” and “objects” and some of these involve non-internal ends that are important reasons, for example when the speaker has an important paternalistic reason which he considers should matter to a reluctant subject, such as in “you ought to take your medicine to save your life”.

The implications of the Theory for the metaethical disputes between realists and expressivists can be illustrated by using an example from Parfit. In his case of a Burning Hotel[41], I see you at the window of a Burning Hotel from which you can only escape by jumping into a canal and I think “you ought to jump.”  Parfit is right that this judgement is not about the subject’s internal motivation and for Parfit this shows that the expressivist tradition errs in not appreciating that important reasons do not have to relate to internal motivation.  Finlay agrees and diagnoses the expressivists failure as due to fixating on normative judgements being about subjective ends and not sufficiently appreciating the possible range of important normative evaluations.  He would explain the judgement in Burning Hotel as being end-relational in having the meaning “you ought to jump for the end of saving your life” with the end omitted by ellipsis as apparent from context, and he would explain that the judgement has the character of being important as it relates to an important end.  In this way, Finlay explains the nature of the normative judgement in Burning Hotel in a straightforward way.

Parfit sees normative judgements as based on normative reasons and assumes that these are basic and cannot be further analysed. But Finlay claims to have met the challenge and analysed normative language. Provided Finlay’s end-relational analysis is sustained, this means that normative judgements can be seen to be on firmer foundations than on Parfit’s approach as their nature is explained (arguably in a rich and elegant way) rather than merely being asserted as being unanalysable.  It also removes a number of concerns with Parfit’s approach, particularly that normative reasons are left as being of a separate sui generis nature, that there is a challenge to explain how we come to know normative reasons, that the cut-off between normative reasons with a separate nature and other purely instrumental reasons appears problematic and that the generally motivating nature of normative reasons needs to be justified.  Although Parfit responds to these challenges, the simpler metaphysics and epistemology uncovered by Finlay’s analysis seems much preferable.

Parfit’s underlying concern is to put on sound foundations our understanding of the nature of normativity – “what it would be for things to matter.”[42]  The end-relational theory can be seen as an advance as it provides a better basis for understanding normativity than Parfit thought would be possible.  It also provides a response to the expressivists by showing that even if normative language is sometimes used pragmatically to express attitudes, its actual meaning is about relationships between “objects” and “ends”.

4.4 Means and Ends

What are the implications of the end-relational view?  I will highlight four points that I personally think are important.

The first is that it changes our appreciation of what goes on in practical reason and morality.  It reveals that all practical thinking has a binary nature, we can look on any discussion and appreciate that ends and means are being considered.  The landscape of practical thinking becomes clearer.

The second implication is that practical reason and morality are shown as being of a straightforward and non-mysterious nature, just being the business of working out what to do by considering ends and means. This dismisses competing views that threaten the purpose of practical thinking and shows that practical reason and morality are fundamentally about reasoning rather than expressing opinions, influencing others, following mind-independent rules or obeying a god. 

The third implication is that ends are shown as important.   The common elliptical omission of ends from our speech reflects our lack of attention to ends as we often take ends as given and concentrate on assessing the best means.  Normativity may use binary object-ends predicates, but if so, the object term normally dominates, and the ends term often has withered to the extent that we are not aware of it.  To deliberate well about how to act should involve consideration of both means and ends and if my ends are unclear and inconsistent, I may act less than optimally.  If we have evolved to think relatively little about ends this is likely to be because this has been an efficient approach.  It will commonly take longer just to properly specify the ends than to decide what to do, so, for much of normative reason, to accept a vague and general end will be an effective heuristic approach.    But such simplistic instincts will not be optimal for many cases, particularly for situations in the interconnected modern world with high stakes and many possible actions. In some cases, such as when making life choices or in politics, we may need to pay more attention to ends than is natural for us.

A fourth, related, implication is to see if we can be more systematic about our ends.  Interest in normative ethics can start from the worry that we make decisions using a rag-bag of ends and a hope to deal with ends more systematically. Our aim may be to work out how to act for the best “all things considered,”  but we will need to debate how far the goal of consistency of ends can be pursued, how far to widen the circle of impersonal concern, and how far we can run beyond our instincts.[43]

5. Conclusion

The starting point of this paper was the thought that some of practical reason outside morality may have a nature that is easier to understand than morality.  It was observed that there is a normative vocabulary based on certain common words including “good” and “ought” which is used fluently throughout practical reason (including morality) and for wider normative discourse.  As argued by Finlay[44], this wide use implies a common underlying semantics so that in all uses, these words concern the evaluation of objects or actions relative to ends, although the semantics may often not be apparent due to ellipsis.  On this end-relational theory of normative language, propositions in practical reason are judgements about relationships between “objects” and “ends” and so have a straightforward metaphysical nature, no more mysterious than judgements identifying facts in the world.  The end-relational theory also succeeds in accounting well for other features of practical reason and appears to be an attractive unifying analysis of practical reason. 

Turning to morality, it is recognised that morality has some special features, but analysis suggests that these do not prevent the end-relational theory applying.  Some people may at times have moral concepts with religious or objectified elements that go beyond an end-relational analysis, but such concepts seem to be erroneous in not reflecting reality and may be avoided by employing end-relational ways of thinking about morality.  Therefore, a naturalistic morality can be seen as having the same simple underlying nature as practical reason.

Notes

[1] (Finlay 2014)

[2] (Schroeder 2016)

[3] (Finlay 2014) and (Finlay 2001)

[4] (Richardson 1997)

[5] Inverted commas are used for these extended meanings of “object” and “end”.

[6] (Finlay 2014, Chapter 6)

[7] Notably, (Hume 1751, 293)

[8] See (Richardson 1997) for a history of the view that it is not possible to reason about ends and a rebuttal of this view

[9] (Finlay 2014, Chapter 3)

[10] (Finlay 2014, Chapter 4)

[11] (Moore 1903)

[12] (Hume 1738, III.i.I), although Hume’s famous passage is more a brief observation than central to Hume’s main views on morality. 

[13] It is assumed that appreciating facts involves conceptual recognition and that a “naïve realism” about facts which believes that facts are purely presented is erroneous.   

[14] Developed from (Finlay 2001)

[15] (Kant 1785)

[16] (Williams 1985)

[17] (Hare 1952)

[18] (Darwell 1998)

[19] (Mackie 1977)

[20] (Darwell 2018)

[21] (Strawson 1962)

[22] For example (Kant 1785), (Sidgwick 1907, 26) and (Mackie 1977, 44).

[23] (Finlay 2014, Chapter 7)

[24] I adopt the “Argument from Queerness” in (Mackie 1977, 38) but following (Kirchin 2012, 102-104) see this as derived from a commitment to naturalism and consider it only applies to limited ways of conceptualising morality.

[25] Developed from (Aitchison 2019)

[26] (Blackford 2019)

[27] (Garner 2007)

[28] (Joyce 2019)

[29] (MacIntyre 1984)

[30] (Pinker 1994)

[31] (Kirchin 2012, 127-138)

[32] (Laskowski 2014)

[33] (Finlay 2014, 195 – 206)

[34] The relativized evaluation is “If one wants to achieve one’s ends, one ought to be practically rational.”  Although in some ways this statement is a tautology, in context it is substantive as it non-trivially assumes that ends and means can be dealt with systematically and that there is a commitment to systematically achieving ends.   

[35] (Finlay 2008)

[36] (Mackie 1977)

[37] (Finlay 2019).  I take (Parfit 2011) as representative of “primitivist moral realism” (“primitivist” as taking normative reasons as basic and indefinable) and (Blackburn 1998) as representative of expressivism.

[38] (Finlay 2006)

[39] Finlay’s view that importance depends on context seems preferable to the view in (Copp 2009) that there are a plurality of different normative systems such as morality, etiquette, law and games.

[40] (Williams 1981)

[41] (Parfit 2011, Part 6)

[42] (Parfit 2011, Volume II, Page 427)

[43] (Sidgwick 1907), (Parfit 1984), (Singer 1981) and (Singer 2005)

[44] (Finlay 2014)

Bibliography

Aitchison, James. 2019. “If Morality Is An Illusion, Does Anything Matter?” MA Philosophy Metaethics Essay.

Blackburn, Simon. 1998. Ruling Passions. Oxford: Oxford University Press.

Blackford, Russell. 2019. “After Such Knowledge – What? Living and Speaking in a World Without Objective Morality.” In The End of Morality: Taking Moral Abolitionism Seriously, by Richard Garner and Richard Joyce. New York: Taylor & Francis.

Chrisman, Matthew. 2014. “Book Review of Stephen Finlay Confusion of Tongues: A Theory of Normative Language.” Notre Dame Philosophical Reviews.

Copp, David. 2009. “Toward a Pluralist and Teleological Theory of Normativity.” Philosophical Issues.

Darwell, Stephen. 2018. “Ethics and Morality.” In The Routledge Handbook of Metaethics, by Tristram McPherson and David Plunkett. New York: Routledge.

—. 1998. Philosophical Ethics. Boulder: Harper Collins.

Finlay, Stephen. 2014. Confusion of Tongues: A Theory Of Normative Language. Oxford: Oxford University Press.

Finlay, Stephen. 2019. “Definining Normativity.” In Dimensions of Normativity: New Essays on Metaethics and Jurisprudence, by D Plunkett, S Shapiro and K Toh. Oxford: Oxford University Press.

Finlay, Stephen. 2008. “The Error in the Error Theory.” Australasian Journal of Philosophy.

Finlay, Stephen. 2006. “The Reasons That Matter.” The Australasian Journal of Philosophy.

Finlay, Stephen. 2001. “What Does Value Matter?” In PhD Thesis. University of Illinois. http://hdl.handle.net/2142/15487.

Fogal, Derek. 2016. “Book Review of Stephen Finlay Confusion of Tongues: A Theory of Normative Language.” Ethics 281-288.

Garner, Richard. 2007. “Abolishing Morality.” Ethical Theory and Moral Practice.

Gregory, Alex. 2015. “Book Review: Confusion of Tongues.” Analysis.

Hare, R M. 1952. The Language of Morals. Oxford: Clarendon Press.

Hume, David. 1738. A Treatise of Human Nature.

—. 1751. An Enquiry Concerning the Principles of Morals.

Joyce, Richard. 2019. “Moral Fictionalism: How to Have Your Cake and Eat It Too.” In The End of Morality: Taking Moral Abolitionism Seriously, by Richard Garner and Richard Joyce. New York: Taylor & Francis.

Kant, Immanuel. 1785. “Groundwork of the Metaphysics of Morals .”

Kirchin, Simon. 2012. Metaethics. London: Palgrave Macmillan.

Laskowski, Nicholas. 2014. “How to Pull a Metaphysical Rabbit out of an End-Relational Semantic Hat.” Res Philosophica 589-607.

MacIntyre, Alistair. 1984. After Virtue: A Study In Moral Theory. London: Duckworth.

Mackie, John. 1977. Ethics: Inventing Right and Wrong. New York: Penguin.

Moore, G. E. 1903. Principia Ethica. Cambridge: Cambridge University Press.

Parfit, Derek. 2011. On What Matters. Oxford: Oxford University Press.

—. 1984. Reasons and Persons. Oxford: Clarendon Press.

Pinker, Steven. 1994. The Language Instinct. New York: Harper.

Richardson, Henry S. 1997. Practical Reasoning about Final Ends. Cambridge: Cambridge University Press.

Schroeder, Mark. 2016. “Value Theory.” The Stanford Encyclopedia of Philosophy. URL = <https://plato.stanford.edu/archives/fall2016/entries/value-theory/>.

Sidgwick, Henry. 1907. The Methods Of Ethics. Vol. 7th Edn. London: Macmillan.

Singer, Peter. 2005. “Ethics and Intuitions.” Journal of Ethics.

—. 1981. The Expanding Circle. Princeton NJ: Princeton University Press.

Strawson, P F. 1962. “Freedom and Resentment.” Proceedings of the British Academy.

Williams, Bernard. 1985. Ethics and the Limits of Philosophy. Cambridge MA: Harvard University Press.

Williams, Bernard. 1981. “Internal and External Reasons.” In Moral Luck, by Bernard Williams. Cambridge: Cambridge University Press.