Classical Muslim and modern historians of Islam have generally recognized the Imām al-Ḥaramayn Abu l-Maʿālī al-Juwaynī (a.d. 1028–85) as an important transitional theologian between the older Ash'arite kalām and the via moderna of Ghazālī and his successors. The Arabic texts of his main surviving works have been published, as well as translations of some into western languages. Careful attention has been given to him in modern studies of kalām such as those of L. Gardet, G. Anawati and M. Allard, although no one has yet written an exclusive monograph of book size on him, as he deserves. This article is devoted to one aspect of his ethical thought, which has escaped treatment until now.

The aspect to be dealt with here is the fundamental question of the nature of value in its application to action. In classical Islam the main issue was whether value terms such as ‘good’, ‘evil’, ‘obligatory’ are definable only by relation to the commands and prohibitions of God, or whether they have objective meanings of their own, which can be applied even to God's actions. This question was much discussed, sometimes as one of the later topics of theology, after the unity and other attributes of God, and sometimes as a foundation for jurisprudence, since the nature of these concepts is basic for understanding how the jurist must describe the sources of Islamic law and how he must construct his theory of legal knowledge.

Rationalism has been vigorous in modern western ethics, from British moralists of the eighteenth century to Kant and his disciples, then to more recent intuitionists such as Moore and Ross. The common elements have been the beliefs that the values of acts and intentions to which we refer by words such as ‘good’ and ‘obligation’ are objective or factual, but in some sense ‘non-natural’, and that they are recognizable by intellectual acts which are different in kind from those by which we recognize characters of sense data or our own mental states. From this brief description we can see that modern rationalism has followed roughly the tradition of Aristotelian ethics. No doubt one could trace its lineage through medieval Christian theories of natural law.

In classical Islam also the existence of rationalist ethics has long been familiar to scholars in the Muʿtazilite theology, with its insistence on the objective nature of good and evil, justice and injustice. Here it was not opposed to naturalistic ethics, as in the modern West, but to the voluntaristic theory developed by the Ashʿarites out of Shafiʿite jurisprudence: the definition of value in action as obedience to divine commands, and the consequent impossibility of knowing right and wrong except from revelation or something dependent on it in one way or another – Traditions, analogy, consensus of the ʿulamāʾ. Ashʿarite ethics have always been well known through the works of Ghazālī, Shahrastānī and others.

The aim of this paper is to present a general review of one theme of Islamic ethical thought that should be of significance and interest to modern philosophers: analytical discussions of the meanings of ethical terms, together with some background information on normative ethics which provide the basis for the evolution of analytical ethics. Two other topics of contemporary interest were also discussed extensively by Muslim thinkers, the psychology of moral action and the question of moral freedom, but I shall leave these aside and concentrate on the first topic.

I shall start with a simplified classification of types of ethical writing in classical Islam, which will provide a frame of reference for the types we are interested in. Two levels of generality in ethical thinking have already been mentioned, ‘normative’ and ‘analytical’. (I should have liked to call the latter ‘philosophical’, but as we shall see there is a theological theory which is consciously antiphilosophical, though just as general.) There are also two traditions of ethical thinking, which we can call ‘secular’ and ‘religious’, according to their conceptions of the proper sources of ethical knowledge. Thus we can derive a fourfold scheme of types of writing on ethics: (A) Normative religious ethics. (B) Normative secular ethics. (C) Ethical analysis in the religious tradition. (D) Ethical analysis by philosophers. I am going to review A, B and D rapidly, and to concentrate on C, where it seems to me that the most interesting thought is to be found.

On several occasions in his writings Ibn Sīnā discusses the problem of destiny (al-qadar), by which he means primarily the problem of reconciling the divine determination of human acts and characters with the rewards and punishments of the after-life, in such a way as to safeguard God's justice to man. This aspect of the problem of theodicy had arisen long before his time out of statements of the Qurʾān and Traditions, and had been settled in their own fashions by Muʿtazilite and predestinarian theologians. Ibn Sīnā as a philosopher could hardly avoid offering a solution, if only to satisfy the doubts of his Muslim public; but he goes beyond a perfunctory answer, and seems to show a genuine interest in finding an intellectually convincing solution consistent with his own philosophy. As would be expected, he interprets the data of the problem in his own terms and comes up with a distinctly Neoplatonic answer, while taking care to express this in a way that might have a chance of acceptance in his religious milieu, Muslim Iran of the early eleventh century a.d.

The word ‘destiny’ in the title and text of this article is used as a deliberately ambiguous translation of Arabic al-qadar. Omitting complexities, we may for our purpose distinguish two main significations of the word in the religious context of earlier Islam. The older usage is ‘predestination’ of human acts and characters by a freely willed decision of God for each person.

The ethical theory of the Muʿtazilites is properly called ‘rationalism’, because it held that the values of human and divine actions are knowable in principle by natural human reason. Since this doctrine and related parts of the theory (mentioned in the next section) were the prevailing ethical theories in the major pre-Islamic religious cultures of Iran and Byzantium, the question easily arises, to what extent were the Muʿtazilites as the first systematic theologians in Islam indebted to these cultures for their ethics?

This question is part of a wider one, that of the origins of systematic theology in Islam (kalām) as a whole. In order to provide a context for my inquiry into the origins of the ethical system, it will be useful to begin with a brief review of the conclusions of scholars on the origins of kalām.

Perhaps it was Maimonides who, more than anyone else, set the tone for early modern views in a famous passage of his Guide of the perplexed:

The subject of deliberation leads us to the core of ethics, because it concerns a method of reaching practical decisions, and to understand this method requires an understanding of how value concepts can be recognized and estimated in practical situations, and this understanding in turn must be based on understanding what value concepts are. In this paper I shall compare the theories of deliberation of two rationalists of different traditions. Aristotle in ancient Greece and ʿAbd al-Jabbār in classical Islam. Any historical link of affiliation between them is either very indirect or nonexistent, and the interest of comparing them does not reside in a search for origins. It consists rather in illustrating two ways in which deliberation has been thought of in the past, and how each way arose out of its own intellectual tradition and environment.

The choice of Aristotle for this study hardly needs explanation, since he alone of the Greeks gave a detailed discussion of deliberation, in his systematic yet probing manner. On the side of Islam the Muʿtazilite school represents the rationalist tradition in theology, and among them only the works of ʿAbd al-Jabbār survive in substantial length. A Persian who lived in Iran and Iraq in the tenth and eleventh centuries (c. 935–1025 a.d.), this theologian wrote his many volumes of the Mughnī and other books in Arabic, as was usual at that time.

Among the debates conducted in Islamic intellectual circles in the early ʿAbbāsid period, one of the most significant was the debate about the nature of value. To simplify the situation a little, we may say that two main theories opposed each other. One was that of the Muʿtazilites, that values such as justice and goodness have a real existence, independent of anyone's will, even God's: this view is classed as ‘objectivism’. The other theory was that of Ashʿarī and his like, that all values are determined by the will of God, who decides what shall be just and so forth: this will be called ‘theistic subjectivism’. Following a struggle between the two doctrines, that of Ashʿarī finally prevailed in most learned circles of classical Sunnite Islam, a result which had far-reaching consequences in law and other spheres of Islamic civilization. As far as the writer is aware, no one has yet examined as a separate problem the reasons why the Ashʿarite theory of value prevailed.

The primary philosophical question about value can be stated broadly thus: What is the common element in all that is called ‘good’, ‘right’, etc.? This question includes the more specific ones of ethics: What constitutes a right action? and, How do we know the right action?

Ibn Rushd (Averroes) devoted most of his efforts as a philosopher to expounding and defending the natural philosophy, psychology and metaphysics of Aristotle, and reconciling with them the doctrines of Islam as he understood them. Reflecting these prominent interests, modern Rushdian scholars have exerted themselves primarily in interpreting his thought in these spheres. Little attention has been given to what he had to say on philosophical questions of value and ethics. Yet it would be surprising if he did not have some well-considered ideas on these questions, in view of his background, education and career on both their Islamic and philosophical sides.

Born into a distinguished family of Malikite lawyers, he must from his earliest years have heard problems of Islamic ethics and law discussed around him in his home. He received a thorough education in Malikite fiqh, and a large part of his paid career in the public service of the Almohad government was spent in appointments to various posts as a qāḍī, including the office of Chief Justice (qāḍị al-jamāʿa) of Córdoba. He also wrote a substantial handbook of Sunnite law, Bidāyat al-mujtahid wa nihāyat al-muqtaṣid. His philosophical education was equally thorough and must have included a study of Plato's Republic and Aristotle's Nicomachean Ethics; and he later wrote commentaries on both these works.

Standards of conduct and character in any society are derived from several sources: from religious prescription, custom, model individuals such as prophets, parents and friends; from literature; finally from the value decisions of everyone in judging his own behaviour and that of others in the past and present, and in prescribing for himself and others in the future. Amid this variety, no accurate answer can be given to the question: to what extent were the ideals of classical Islamic society derived from specifically Islamic sources – the Qurʾān and Traditions as interpreted by Muslim scholars? But at least it is certain that these sources were of predominant importance in all Muslim countries until the nineteenth century and still prevail in the more conservative countries and in all rural areas. This was and is inevitable wherever the basic education of children is the study of the Qurʾān, with its well defined ethical precepts and attitudes to life on earth.

So it will be worth our while to review briefly the norms emphasized by the scripture of Islam, as a prelude to the main topic of this paper, which will be the methods by which norms have been thought to be properly learned by the Muslim community; in other words the methods of ethics according to Muslim thinkers.

A considerable debate was carried on in classical Islam over two central and related questions of ethics.

A. A question of ontology: What is the nature of ethical value concepts such as the good and the just?

B. A question of epistemology: How can man know the presence and force of these concepts in particular situations? These are philosophical questions, and to some extent they were argued in terms of truths determinable directly by human intellect. But the debate was mainly conducted by Muslim theologians and jurists who did not always distinguish direct arguments from arguments on the corresponding questions of scriptural interpretation: what does the Qurʾān teach on the nature and knowledge of ethical value concepts? The present article attempts to answer the latter questions only, by a study of the Qurʾān in its own context.

As a framework for study it will be useful to set forth in schematic form those answers to the two questions which may be considered historically possible in the age of the Qurʾān, leaving out many modern ethical theories which do not satisfy this condition. (‘Right’ will be taken here as the standard concept, but we shall be dealing also with ‘just’, ‘good’, ‘wrong’, ‘unjust’, ‘evil’.)

This book is based on articles, published in orientalist journals during the last twenty-five years and recently revised to make a consistent book and bring it up to date in a few places.

The first acknowledgement I should like to make is to my wife for her patience and her encouragement of my writing over such a long period.

Next, I wish to thank the institutions which have financed the free time which made these studies possible: namely, the University of Michigan and the State University of New York for sabbatical leaves and various research grants, as well as the Ford Foundation and the Simon E. Guggenheim Foundation which allowed me to extend sabbaticals on two occasions.

Thirdly, I acknowledge permissions granted to republish in book form the articles held in copyright by the following publishers: Scribner and Sons for articles in the Dictionary of the Middle Ages (2); the SUNY Press (Albany) (2); The Muslim World (3); Éditions de l'Institut Supérieur de Philosophic (Louvain); International Journal of Middle East Studies; Bruno Cassirer (Oxford); Caravan Press (Delmar, New York); Journal of the American Oriental Society; Studia Islamica (2); and Bulletin of the School of Oriental and African Studies (London). Full references to the original articles are given after this Preface, showing slight variations in the titles of three articles.

Fourthly, I am grateful to several secretaries at the University of Michigan and the State University of New York at Buffalo for their work in typing these articles.

Islamic ethics should be of interest to Muslims and non-Muslims alike in two aspects. The first is its central place as the core of Islam, if we can include Islamic law as an integral part of ethics. The Qurʾān repeatedly uses the phrase ‘those who believe [in God] and do good works’, taking it for granted that these two attributes belong to the same group in extension, and that the first is a prerequisite of the second – but also that the first would be insincere and not true belief without the second. Following this lead, the legal profession in the first two centuries of Islam tried to make the law of the sharīʿca cover every ethical situation and to make the study of this law the culminating study in Islamic education. Thus, since Islamic education was the most formative element in Islamic civilization, the important rôle of ethics in this civilization becomes obvious.

The second aspect of interest lies in the theological and philosophical theories that were constructed to support normative ethics. These theories all belong to the main Western lines of rational theology and philosophy and have little relation to the theologies and philosophies of East Asia. This relation to Western intellectual history has, unfortunately, rarely been recognized in the West. As a result, Islamic thought has been generally neglected and ignored in Western histories of theology and philosophy.

Al-Ashʿarī, Abu l-Ḥasan ʿAlī b. Ismāʿīl (approximately 873–935), a leading conservative theologian in Sunnite Islam, after whom an enduring school of theology is named.

Born in Baṣra, he received the usual Islamic education in the Qurʾān, Traditions, Arabic philology and sharīʿa law, then studied theology (kalām) under the foremost Muʿtazilite theologian of the Baṣra school, Jubbāʾī (d. 915). He gained a reputation as an excellent debater on theology in the mosques and wrote works in the rationalist tradition of his master, which have not survived. At the age of forty, however, he changed his doctrinal position to a more traditionalist one and became a strong opponent of the Muʿtazilites. This event is reported as resulting from three dreams, in the first two of which Muḥammad commanded him to defend a more traditional Islam, while in the third he insisted that he should not abandon the dialectical method of kalām but should use it to combat Muʿtazilite rationalism. Then for the rest of his life Ashʿarī championed traditional Islamic theology, moving at some time from Basra to Baghdād where he died. Among his few surviving works from this later period are two that present in dialectical form his doctrinal positions and arguments, the Ibāna and the Lumʿa, as well as a valuable survey of the sects of Islam, Maqālāt al-Islāmiyyīn.

The most distinctive and influential feature of Ashʿarīs new ‘conservative’ theology is its method.

I would like to explore in this paper what is involved in the notion of a self, of a responsible human agent. What is it that we attribute to ourselves as human agents which we would not attribute to animals?

This question takes us very far indeed, and into several issues of capital importance in philosophy. I am not even going to try to sound them all. But I'd like to make a preliminary exploration of the terrain, using as my guide a key notion which has been introduced recently by Harry Frankfurt, in order to see how well the territory of the self may be mapped with its aid.

The key notion is the distinction between first- and second-order desires which Frankfurt makes in his ‘Freedom of the will and the concept of a person’. I can be said to have a second-order desire when I have a desire whose object is my having a certain (first-order) desire. The intuition underlying Frankfurt's introduction of this notion is that it is essential to the characterization of a human agent or person, that is to the demarcation of human agents from other kinds of agent. As he puts it,

Psychology is a vast and ramified discipline. It contains many mansions. But this does not prevent it from being intellectually divided against itself. I would like to discuss in particular one major division in psychology, and in so doing suggest a way of reconciling the two sides. Out of such reconciliation, there might come a new perspective on the discipline.

The deep division in the discipline which I would like to examine concerns basic notions about science and the methods of science. At bottom it is a difference of epistemology. The epistemology which is dominant among experimental psychologists in America is the one which emerged with the scientific revolution in the seventeenth century. There are many ways of characterizing this view of science, but for the purposes of this discussion we can single out two features, because they are the principal sticking points in any confrontation between the two epistemologies.

The first is the principle that science must be grounded in data that are intersubjectively univocal. This point is often put by saying that experiments must be replicable by anyone, and any results must be such that others can check them. Science achieves its impressive show of objectivity through intersubjective agreement. Now this puts a restriction on what can be counted as data of a science. There are many ranges of judgements we make in ordinary life which cannot come close to meeting this requirement of intersubjective univocity.