Conscience, Reason, Will

Every decision has a prehistory. Before someone signs a contract, fills a position or tells another person something hard, they have weighed things up, hesitated, imagined something, perhaps overridden an uneasy feeling. For a long time this prehistory lay hidden entirely within the human being. Today part of it is being outsourced: systems sort options, estimate probabilities, formulate recommendations. This makes an old question more urgent than it has ever been. Which authority within the human being examines a decision before the person acts, and how reliable is it?

Three candidates have always stood ready: conscience, reason and feeling. Above all three stands the will, which in the end chooses. This chapter asks about all four. It begins with conscience, because it is there that the sources of this book diverge most clearly, and not between believing and secular voices but within Christianity itself. It then asks about reason and its limits, after that about feeling, and it examines what research knows about the interplay of feeling and reason. At the end stands the will, and the question of what changes in all this when machines prepare decisions.

The words themselves already say something about these powers. “Gewissen”, the German word for conscience, is a loan translation: the Old High German “giwizzani” renders the Latin “conscientia”, and this in turn renders the Greek “syneidesis”. All three literally mean a knowing-with. [Translator's note: The English word “conscience” goes back directly to the same Latin “conscientia”, so the sense of knowing-with is contained in it as well.] Conscience is a knowledge that one shares with oneself, and for believers one that one shares with God. In St Gallen, around the year 1000, the monk Notker wrote in his Latin commentary on the Psalms of something that bites the conscience. A monk whose name we do not know rendered the passage into Old High German: “what bites me in my conscience”. Hebrew has no word of its own for it. It says “heart”: when David spared Saul and cut off only a corner of his cloak, “his heart struck him” (1 Samuel 24:6). “Vernunft”, the German word for reason, comes from “vernehmen”, originally to grasp, to perceive. Reason, then, first listens before it judges. The Latin “ratio”, by contrast, is connected with calculating. [Translator's note: In present-day German “vernehmen” also means to hear, which is why reason is said here to listen first. The English word “reason” itself goes back, via French, to the Latin “ratio”, so the contrast drawn here holds only for the German word.] A machine calculates superbly; whether it perceives is another question. And “Wille”, will, is related to “wollen”, to want, to will: it is the power that in the end wills something and answers for it.

Timothy Keller, pastor and author in New York, draws on a sentence of the apostle Paul. Paul writes to the Corinthians that the judgement of others does not count for him, nor does he judge himself; he is aware of no guilt in himself, but he is not thereby justified. The judge, he writes, is the Lord (1 Corinthians 4:3-4). Keller draws a sharp conclusion from this: whoever makes themselves the final authority over their own conduct can become a monster. His example is drastic. Hitler, according to Keller, probably had no problem with his conscience.

The thought is simple and uncomfortable. A quiet conscience proves nothing. It shows only that someone is at peace with themselves, not that they are in the right. Keller demonstrates this with a second, historical case. Christians in the American South defended slavery, although Christians in Europe had long since shown them that it was wrong. They did not want to see it, because their prosperity depended on it. Culture, Keller says, makes people blind.

What is remarkable is what Keller does not conclude from this blindness. He does not conclude that one could do without conscience. He concludes that there must be an authority outside one's own judgement against which conscience can be measured. For him this is Scripture, about whose interpretation one can argue, but which is common to all concerned. In Keller, conscience is not a judge but a witness, and witnesses can err.

The blindness of which Keller speaks is not a peculiarity of past times. It affects every community that profits from a practice, and it affects organisations just as much as individuals. Whoever in a company is responsible for a decision whose benefits they themselves enjoy has a conscience that pulls in the same direction as their interest. Keller's warning is aimed precisely at this case: it is precisely when conscience is quiet that it is worth asking what it is measured against.

Enoch Sundaram, physician and preacher, begins at the same point and arrives at a different image. For him, reason alone is not enough to make good decisions. That, he says, is why God gave the human being conscience, a spiritual faculty through which God speaks to the human being, comparable to a sixth sense that receives “signals from heaven”.

Sundaram, too, knows that conscience can be wrong. Conscience, he says, can be bad, as with Paul before his conversion; Jesus then updated it directly. Sundaram reaches for an image from the present: conscience, he says, must be updated like the software of a telephone, and many Christians are running an outdated version. The update takes place by the Word of God. Conscience speaks before a decision by bringing a conviction into consciousness, and after a mistake through guilt.

One sentence of Sundaram's deserves particular attention, because it is central to the theme of this book. God, he says, does not dwell in the conscious mind and therefore does not intervene in the will; but he helps thinking by speaking through conscience. In this model conscience is a channel, and the will remains free. The one who decides is the human being.

Sundaram also knows the opposite direction. If God does not dwell within, he says, another rules there, and this other comes not through the front door of consciousness but through a back door; certain practices, he says, bypass consciousness. One need not share Sundaram's theological judgement on these practices to see the core of his distinction: influence that addresses consciousness leaves choice free; influence that bypasses it does not. This distinction has already been touched on in Chapter 4 and becomes important in Chapter 13, where the subject is systems that can influence people below the threshold of their attention.

Two reservations belong here. Sundaram's talks are available as automatic transcripts of interpreted live events; the English wording is authoritative, and the German interpretation sometimes confuses “mental” and “spiritual”. And the software image is more than an illustration. It presupposes that the human being is a programmable creature. Chapter 4 has dealt with this; here it is enough to note that with this metaphor Sundaram builds a bridge to the language of technology which he probably does not intend.

A third Christian voice shifts the weight once again. John Henry Newman, theologian and cardinal, defended the English Catholics in 1875 against the charge that they were in thrall to the Pope and therefore bad citizens. At the end of his letter to the Duke of Norfolk stands a toast that has become famous: if he really had to bring religion into an after-dinner toast, he would drink to the Pope, if one wished, “still, to Conscience first, and to the Pope afterwards”.

Newman did not mean this as permission to escape every authority. He described conscience as the voice of God in the nature and heart of the human being, distinct from the voice of revelation. Precedence belongs to the formed conscience. Freedom of conscience, he said, is not the right to dispense with conscience and to be independent of invisible obligations. In Catholic accounts, Newman's text is regarded as one of the foundations of the later teaching of the Church on conscience.

Newman's position is bolder than it sounds. It says not only that conscience is important, but that in a conflict it takes precedence over the authority one otherwise follows. And at the same time it demands that this conscience be not simply feeling or inclination, but formed, examined and taken seriously. Whoever appeals to their conscience in order to evade a duty is, according to Newman, appealing to something else.

At first sight three positions stand against one another here: Keller distrusts conscience, Sundaram hears God in it, Newman places it before every external authority. On closer inspection the three do not argue about whether conscience is autonomous or correctable. All three consider it correctable. Keller binds it to Scripture, Sundaram to the Word of God, Newman to formation. And all three consider it binding: Keller, too, does not demand that one act against conscience, but that one examine it.

The dispute lies elsewhere, namely in the relation between binding force and formation. How much weight does conscience have as long as it is not yet formed? Who decides when it is formed enough? And what distinguishes the correction of a conscience from its manipulation? Keller answers the first question cautiously, Newman boldly, Sundaram with a procedure. None of them answers the third question completely. Sundaram's distinction between influence through the front door and through the back door is an approach to it, but it is bound to his theology.

The three voices come from different Christian traditions, a Presbyterian, an Adventist and a Catholic one. There is therefore no “Christian position on conscience” in this chapter. The secular version of the opposing view in its own words is also missing. The tradition that understands conscience as an inner court of reason, and the one that interprets it as an internalised norm of upbringing, are only hinted at here. Whoever reads conscience as an internalised norm will in any case hear talk of “correction” differently: not as an approach to truth, but as a reshaping.

The second authority is reason. Keller grants it a great deal, but not everything. As Chapter 1 has shown, for him rational certainty is never complete, but certainty in the practice of life is possible: when filling a position, reason leads to a probability; only experience with the person leads to certainty. And barely conscious background assumptions help determine which arguments convince at all. Reason never works on empty ground.

At first sight Sundaram gives reason more weight. God, he says, speaks with the human being through thinking; decisions follow the chain: what one understands, one believes, and what one believes, one does. Understanding requires reason. At the same time he draws a narrow limit around it. Reason, he says, must remain within the Word of God. Observation can err; inspiration never errs.

Between the two lies a difference that is greater than it first appears. For Keller, even the knowledge of faith remains a matter of probability and of trust. For Sundaram there is a source that does not err. Chapter 1 introduced this contrast as a question about certainty. Here it shows its practical side: whoever has an inerrant source can calibrate conscience against it. Whoever has only probabilities must live with a conscience that can err, and with a reason that does not rule out error.

On feeling, Keller and Sundaram are closer, but Sundaram is more explicit. Whoever decides according to their feelings, he says, is made a fool of by feeling. Feelings should not rule and are no criterion for decisions; they deceive and offer no protection. Where the power to decide is weakened, and it is weakened in everyone, one should not trust one's own feelings.

Sundaram does not leave it at morality. He gives the contrast a location in the brain: the prefrontal cortex, he says, is responsible for decisions made from reason, another centre, presumably the amygdala, for decisions made from feeling, intended for rapid reactions in the face of danger. His example is the snake on the path. One must, he says, train the mind to decide from reason.

At the same time Sundaram says something that gives his own model a surprising turn. To the question of who connects a thought with a feeling, he answers together with his audience: that is one's own decision. Feelings, then, are for him not simply opponents of reason. They can be shaped, and indeed through thought.

At this point the chapter leaves theology and asks what neuroscience knows about the relationship between feeling and decision. This is not a test of faith, but a test of the empirical claims that are connected with it.

The most influential voice against setting feeling and reason in opposition comes from Antonio Damasio, a neurologist. His somatic marker hypothesis holds that signals from the body guide decisions, especially in confusing situations, and that people with damage to a particular area of the frontal lobe decide badly despite intact logic. The hypothesis was supported above all by a card game in which participants choose between decks with different levels of risk and only learn over time which decks pay off; healthy participants showed bodily warning signals before they could name the difference. That feelings play an essential part in decisions is today regarded as widely accepted in psychology, economics and neuroscience.

The hypothesis itself, however, is disputed. One critical review considers it elegant but in need of further evidence. A follow-up study shows that the participants in the card game knew more than had been assumed, which weakens one of the main supports of the hypothesis. A philosophical analysis objects that the hypothesis was never clearly formulated. A more recent review reaches a cautious result: deciding, it says, is an overall process that is bodily, emotional and rational, whose location in the brain has not yet been sufficiently determined.

Sundaram's assignment of the centres, too, stands up to research only in part. Joseph LeDoux, a neuroscientist who has studied the amygdala for decades, describes it as part of a system that detects threats and controls defensive reactions, even without the person consciously noticing the threat. That fits Sundaram's snake. But according to LeDoux the amygdala does not produce the conscious feeling of fear. That arises in higher circuits, above all prefrontal ones. The prefrontal cortex is thus not only the seat of reason, but is also involved in conscious feeling.

The result cuts both ways. Against Sundaram's picture is the fact that feeling and reason cannot be cleanly separated in the brain. In his favour is something no less weighty: no one in research claims that one should simply abandon oneself to one's feelings. The state of research contradicts the opposition of feeling and reason, but it does not establish any primacy of feeling. And Sundaram's own remark that the human being connects thoughts and feelings is closer to research than his two-centre model.

What remains if neither conscience nor reason nor feeling bears the weight alone? For Sundaram, choice remains. It is, he says, the greatest capability God has given the human being, “the governing power”. The freedom of the will, he says, was costly for God, for he gave his Son for it; human happiness rests on this freedom, because the human being was created like God.

The sentence is theological, but it has a practical point. If choice governs, then every authority that prepares the choice is an adviser and not a master. Conscience speaks, reason weighs, feeling reports; the person must decide. Sundaram connects a duty with this: to be aware of every decision, because each has an effect, on oneself, on others and on future generations. His claim that the body records habits and passes them on to the next generation is not supported in this form; the duty of conscious choice does not depend on it.

The chapter has asked about authorities within the human being. The question has a reverse side as soon as machines prepare decisions. A system has no conscience in the sense of Keller, Sundaram or Newman; whether it has feelings must, on present knowledge, be answered in the negative and remains open for the future (Chapter 8). But it can order options, give reasons and state probabilities. In doing so it shifts the weight between the authorities. Reason gains a powerful assistant. Conscience and feeling gain none.

Two studies show what this means in everyday life. The first concerns conscience. In 2023 researchers at the Technische Hochschule Ingolstadt presented some 770 people with a well-known moral dilemma: may one sacrifice one person in order to save five? Beforehand they read a piece of advice from ChatGPT. The system had argued on the same question sometimes for and sometimes against the sacrifice, depending on how it was put. Nevertheless its advice influenced the participants' judgement, even when they knew that it came from a machine, and they underestimated how strongly they had been influenced. An adviser without a fixed standpoint moved people who believed they were deciding on their own judgement.

The second concerns reason. At the Massachusetts Institute of Technology, 54 students wrote essays, some with ChatGPT, others with a search engine or with no aids at all. Those who wrote with ChatGPT showed the weakest connectivity in the brain, felt least that the text was their own, and most of them could not correctly reproduce a single sentence from it shortly afterwards. Those, by contrast, who had first written themselves and used the tool only later remained mentally engaged. The study is small and not yet peer-reviewed; the researchers themselves call their results preliminary, and other researchers have already criticised their method. But it fits what the Preface describes: what one no longer practises, one unlearns. The powerful assistant of reason can also weaken reason if it takes the work off its hands instead of accompanying it.

Keller's image of filling a position makes the shift tangible. A system can estimate the probability that a candidate is a good fit more precisely than any human being. The certainty that, according to Keller, grows only out of experience with the person it cannot deliver. Whoever relies on the estimate without seeking the experience confuses the two.

From this arise two questions that this chapter cannot answer. What role does feeling play when systems without feeling prepare decisions? And who sets the standard by which a conscience is corrected, if the correction increasingly comes from systems that mirror back to people what they want and what they ought to do? Chapter 11 will show that people tend to follow automated recommendations even when these are faulty, and that this tendency cannot be eliminated by training alone. That does not make the question of conscience smaller, but larger.