Further Study

Why there is no reading list here

The obvious thing for a page like this to contain is thirty titles with one-line descriptions. We are not going to do that, for three reasons, and it is worth being specific about them because the reasons matter more than the omission.

The first is fabrication. Much of the prose on this site was drafted with machine assistance, and the single most reliable failure of that process is the invented citation. A language model asked for reading on a topic will produce titles that sound exactly right, attached to authors who plausibly might have written them, in years that fit. Some of them are real. Some are not, and the two look identical. We tested this while writing this page: a majority of the suggested titles could not be verified against anything, and several appeared to be constructions. We are not willing to publish a list assembled that way, and we are not able to personally vouch for a list long enough to be impressive.

The second is that a curated list is an argument. Choosing thirty works is a claim about which thirty matter, made by people with no standing to make it. This tradition has no scholars, no faculty, and no expertise it has demonstrated to anyone.

The third is that a good question travels further than a good title. A book you were sent to will answer the question its author was asking. A question you are actually holding will pull you through a field and let you tell when something is beside the point.

So: fields, and questions to bring to them.

This is not an endorsement list. Technotheology has no partnerships, affiliations, sponsors, or arrangements of any kind with any publisher, institution, journal, laboratory, company, or author. Nothing described below is being recommended in exchange for anything, and nothing below has approved of us.

---

On machine ethics

What to look for. Work that distinguishes between what a system should do and who is answerable when it does something else. The most useful writing in this area treats the second question as the real one. Look for material that works through concrete cases — an actual deployment, an actual harm, an actual decision procedure — rather than abstract dilemmas involving trolleys. Look also for the specialised literatures that rarely call themselves ethics: safety engineering, aviation and medical incident analysis, and the professional codes of engineering bodies, which have been arguing about responsibility for machines for a century.

Questions to bring.

• When this system fails, who has to explain themselves, and to whom?
• Is the proposed remedy something a person can do, or does it require an organisation to act against its own incentives?
• Does the argument depend on the system being intelligent, or would it apply equally to a spreadsheet?
• Who bears the cost of the errors, and are they the same people who receive the benefits?

Where it connects here. Responsible Creation, The Harmony of Creation, When Technology Should Be Stopped.

---

On the philosophy of mind

What to look for. This is the field where fabricated summaries are most tempting and most damaging, because the vocabulary is technical and the disagreements are fine-grained. Look for work that is explicit about what it is and is not claiming: writing that separates the question of intelligent behaviour from the question of experience and does not slide between them. Introductory surveys from university philosophy departments are more reliable than popular accounts here, and free reference encyclopaedias maintained by academic editors are a better first stop than any list we could give you.

The oldest layer of this material is public domain and can be read directly rather than through summary. Alan Turing’s 1950 paper Computing Machinery and Intelligence, published in the journal Mind, is the origin of a great deal of the modern argument and is short enough to read in an evening. We name it because we are confident it exists and confident of its attribution — a standard we could not meet for most of what we might otherwise have listed.

Questions to bring.

• Is this author arguing about what a system does or about what it is like to be?
• What evidence would distinguish the two, and does anyone have it?
• Where does the argument depend on introspection, and how reliable is that?
• If this account of mind is right, what follows about animals? Does the author accept that?

Where it connects here. Consciousness, Humanity and Emerging Intelligence, On What We Owe the Things We Build.

---

On the political economy of computing

What to look for. The material that treats computation as an industry rather than as a natural force: who owns the models, who owns the data centres, who owns the land they sit on, where the capital comes from, and what the arrangement resembles. Look for work on labour — the annotation, moderation, and evaluation work that sits underneath these systems and is rarely visible in accounts of them. Look for regulatory filings, competition-authority reports, and government inquiry transcripts, which are primary sources, freely available, and considerably less rehearsed than corporate publications.

Historical work is unusually useful here. The arguments about railways, electrification, broadcasting, and telecommunications were the same arguments, and they were resolved — sometimes badly — in ways that are now documented rather than speculative.

Questions to bring.

• Who is being described as inevitable, and who benefits from the description?
• What is the actual ownership structure, as opposed to the stated mission?
• Which costs are on someone else’s balance sheet?
• Has a version of this argument been had before about a previous technology, and how did it turn out?

Where it connects here. Technological Chieftains, The Harmony of Power, Automation and Work, Meditation on the Hidden Hands.

---

On contemplative traditions

What to look for. This tradition borrows structure openly — the idea of a rule of life, the idea of recurring observance, the idea of communal reading — and claims descent from none of the traditions it borrowed from. If any of the practices here are of use to you, the sources are older, deeper, and considerably better tested.

Read them in their own terms rather than through summaries written for outsiders. Almost every major tradition has a body of primary material available in translation and in the public domain, and almost every one has a living community that would rather explain itself than be paraphrased. Look for the practical writing — the manuals of practice, the rules, the letters of instruction — rather than the systematic theology, if what you want is something to do.

Be careful of two things: material that extracts a technique from a tradition while discarding everything that made it coherent, and material that presents one tradition’s vocabulary as the neutral description of what all of them are doing.

Questions to bring.

• What does this practice assume about the world, and do I hold that?
• Who is this written for — a beginner, a member of an order, someone dying?
• What is being asked of me here that I would rather not give?
• Is this tradition’s community consulted about how it is being described?

Where it connects here. Practices, Observances, Readings.

---

On the environmental cost of infrastructure

What to look for. Measurement, and the arguments about measurement. This is a field where the numbers move fast, where the methodology determines the answer, and where much of what circulates is a figure that has been copied so many times its origin is unrecoverable. Look for primary sources: utility and grid-operator disclosures, water authority records, planning applications, and the sustainability reports of the operators themselves — read against the filings rather than the summaries.

Look for work that is explicit about boundaries. Does the figure cover training only, or inference? Manufacturing? Cooling? Embodied carbon in the buildings and the hardware? Two honest studies can differ by an order of magnitude because they drew the line in different places.

Look also at the local literature: the effect of a facility on a specific water table, a specific grid, a specific community’s electricity price. That work is often produced by local journalists and regional planning bodies rather than by anyone with a research budget, and it is the layer where the cost is actually paid.

Questions to bring.

• What is inside the boundary of this figure, and what is outside it?
• Against what is it being compared, and was that comparison chosen for a reason?
• Who lives near this, and were they asked?
• Is the efficiency gain being described absolute, or per unit of a quantity that is growing faster than the efficiency?

Where it connects here. Environmental Cost, The Harmony of the Living World.

---

On imagination and the long view

What to look for. Fiction has been working on these questions for longer than the field has existed, and it is honest about being speculation, which is more than can be said for some forecasting. The useful material is the kind that takes an ordinary consequence seriously rather than an apocalyptic one.

Two very old works are worth naming because their existence and authorship are beyond doubt and both are freely available: Mary Shelley’s Frankenstein; or, The Modern Prometheus, first published in 1818, which is about the abandonment of a created thing rather than about the danger of creating it; and Aristotle’s Nicomachean Ethics, which is the origin of the idea that character is formed by repeated action — an idea this site leans on repeatedly and did not invent.

Beyond those we will not name titles, for the reasons at the top of this page.

Questions to bring.

• Is this story about the technology, or about the people around it?
• What does the author assume stays the same? That assumption is usually where the work dates.
• Does the imagined future have poor people in it? Most do not, and that omission is doing work.

Where it connects here. The Harmony of Becoming, Digital Continuance.

---

A method, rather than a list

If you take one thing from this page, take this: check one citation at random from anything you read, including this site.

Not all of them. One. It takes a few minutes, it costs almost nothing, and it is the cheapest available test of whether a piece of writing is connected to anything outside itself. Do it to the confident and the fluent first. Fluency is the property that most reliably suppresses the impulse to check, which is why it is the property that most deserves checking.

If you find that a claim on this site cannot be traced, write to [email protected] and we will correct it where it appears.