Naming the Shadows: Literal Translation and Plato's Cave

In Republic Book VII, Plato has Socrates describe the prisoners in the cave naming the shadows they see (515b–515e). The prisoners take the shadows to be the whole of reality, because the shadows are the only thing they have ever perceived. Naming, in this passage, is what gives the shadows their value: by calling them by name, the prisoners treat the shadows as though they were the things themselves, rather than as poor copies of things carried behind a wall they cannot see. This essay argues that GCSE teaching of literal translation works in the same way. Students name the literal translation, that is, they accept it as accurate, because the literal translation is the only thing they know. They have no statue to compare it against, only the shadow itself, and so the shadow becomes, for them, indistinguishable from the truth.

The argument proceeds in four stages. First, I build the allegory itself, mapping each element of the cave onto the practice of literal translation in the classroom. Second, I extend Plato's own theory of Forms into the domain of language, since this extension is what allows the allegory to answer one of its strongest objections. Third, I set out four objections to the allegory, taken as seriously as I can put them. Fourth, I answer each objection in turn, and show that the allegory, properly understood, survives all four.

Building the Allegory

The prisoners are named after the shadows because the shadows are the only thing they know. Socrates says of them that “the truth would be literally nothing but the shadows of the images” (515c). This is not a claim about what the prisoners believe carelessly, as though they might know better and simply fail to notice. It is a claim about the limit of their experience. They have no access to anything else, so the shadow is not mistaken for the truth in the sense of an error that could in principle be corrected by looking more carefully. The shadow simply is the truth, as far as they are concerned, because nothing in their experience could ever indicate otherwise.

Students learning GCSE literal translation are in the same position, and the parallel is worth drawing out carefully rather than merely asserted. They name the literal translation, that is, they accept it, because the literal translation is the only thing they know. The Greek term for truth, ἀληθές (to alēthes), captures what is at stake: for the prisoner, and for the student, what passes for ἀληθές is simply whatever they have been given, with no comparison available against which it could be judged partial or misleading. The student is not comparing the literal translation against other translations, and they are not comparing it against the original language, because they do not have access to either. They sit an exam in which the literal translation is treated as the object of study in its own right, not as a stepping stone toward the original, and so the translation absorbs all the authority that would otherwise belong to the source text.

As I have argued elsewhere, Voltaire wrote that “woe to the makers of literal translations, who by rendering every word weaken the meaning,” and went on to say that in doing so, the letter kills while the spirit gives life. Translations of this kind can be misleading even when nobody involved intends to mislead anyone, and even when the translation itself is competently made. The danger is not that the translator has been careless. The danger is structural: a single rendering, repeated often enough and unchallenged by any alternative, comes to occupy the entire space in a student's mind that the original should have occupied. To summarise the premise before moving to the allegory proper: students learn translations that they take to be fully accurate, because accuracy is not something they are in any position to assess. They are not comparing translations against each other, and they do not experience the original language, so they are governed entirely by the single translation they are given, in precisely the way that the prisoners are governed entirely by the single shadow in front of them.

The mapping onto the allegory can now be set out directly, element by element. The people are the students. The shadows are the literal translations. The statues carried behind the wall, out of the prisoners' sight, are the original texts, in Greek or Latin, that the students never encounter directly. The fire that casts the shadows is the pedagogical apparatus itself, the exam board, the textbook, the scheme of work, whatever produces the translation and puts it in front of the student as the object of study. The pain of the prisoner who is dragged out of the cave and up toward the light is the pain of learning the original language properly: the disorientation of discovering that the thing one took to be solid and whole was only ever a flattened outline of something with far more depth and texture than the shadow could show.

The Forms Extended to Language

Plato's theory of Forms gives a second, related image that strengthens the allegory considerably, and which I want to bring in explicitly rather than leave implicit. In Republic Book X, Plato has Socrates distinguish three beds: the Form of the bed, which is the true and single reality of what a bed is; the carpenter's bed, which is one remove from the Form, an imperfect physical instance of it; and the painter's bed, which is a further remove again, since the painter does not work from the Form but merely copies the appearance of the carpenter's already-imperfect bed (596a–598b). The painter is two removes from the truth, not one, because the painter copies a copy.

This structure maps onto language directly, and this mapping is the key to answering one of the objections raised below. A thought, in the mind of the person who first has it, is analogous to the Form: it is the single, complete thing that everything else is a version of. When that thought is put into language for the first time, in the original text, it undergoes one remove: thought becomes language, in the way that the Form becomes the carpenter's bed. This is already an imperfect step, since no sentence ever captures a thought with total fidelity, but it is only one step.

Translation does not repeat this single step. It requires two. First, the translator must reconstruct the original thought from the original language, converting language back into thought, which is itself a further act of interpretation rather than a neutral retrieval. Second, the translator must then put that reconstructed thought into a new language, converting thought into language a second time. The sequence for the original text is thought, then language. The sequence for the translation is thought, then language, then thought again, then language again. Translation is not merely twice removed from the original text in some loose or metaphorical sense. It is twice removed from the original thought in the same precise sense that the painter's bed is twice removed from the Form: it is a copy of a copy, not because translators are careless, but because the entire process necessarily runs through the thought-to-language step twice rather than once.

Objections to the Allegory

Having built the allegory and extended it through the theory of Forms, I want to argue against it as forcefully as I can. There are four objections worth taking seriously, and I will state each one in its strongest form before answering it in the next section.

Objection 1

First, what is truth? Literal translation preserves some meaning; it is not a mere shadow with nothing of the original left in it, in the way that a shadow retains only outline and loses all colour, texture, and depth. Translation is twice removed from the original, in the sense that Plato uses of art and representation, but so are many other things we do not treat as shadows and do not subject to the same suspicion. Translation is also an active process, carried out by a translator making deliberate, considered choices, not a passive distortion that happens automatically to the original as it passes through a medium the way light passes through water and bends.

Objection 2

Second, literal translations preserve features of the original such as roughness, repetition, and strangeness of syntax, precisely the features that idiomatic translations smooth away in the name of readability. If literal translation retains these features, then we cannot cleanly say that literal translations are shadows while idiomatic translations are truth, since the allegory as I have built it seems to imply exactly that. The allegory implies a single scale running from shadow to truth, with literal translation closer to the shadow end, but literal and idiomatic translations may simply preserve different things, not more or less of the same thing, which would make the whole scale a false one.

Objection 3

Third, translation is thought to be twice removed from the original, but so is paraphrase, so is memory, so, in a sense, is the original text itself once it is read and interpreted by anyone other than its author. If everything is removed from some prior thing, whether by two steps or by some other number, it is not clear why translation alone should be singled out for this particular kind of scrutiny, rather than treated as one instance among many of a perfectly ordinary and unavoidable fact about how meaning travels between minds.

Objection 4

Fourth, the turning of an object into a shadow is a passive process in the allegory: the object is simply held up in front of the fire, and the shadow falls where it falls, without anyone needing to make a judgement about how the shadow should look. Translation is not like this. It is an active and careful process, undertaken by a translator exercising sustained judgement over years of training, not a mechanical projection that happens automatically without anyone choosing anything at any point.

Responses to the Objections

Each of these four objections can be answered, and answering them properly sharpens the allegory rather than forcing me to abandon it.

Response 1

To the first objection: there are multiple translations and multiple interpretations of any original, so a translation is not a single statue that turns into a single shadow, because sometimes there is no single statue to begin with, only a cluster of plausible statues, each casting a slightly different shadow. This looks like a strong objection, but it can be answered without weakening the allegory. Translation is an act that removes the very things that make up the fullness of the original: its context, its idiom, its historical resonance, the associations a word carries for a native speaker but not for a learner. The statue is the entire source, context included, and it is precisely that context which gets stripped away in the crossing into the shadow. That there may be several statues carved from the same block, representing several plausible translations of an ambiguous original, does not change the fact that each one casts its own shadow by exactly the same process of removal. Multiplicity of statues does not rescue any single shadow from being a shadow.

Response 2

To the second objection: literal translations do preserve roughness and repetition, but they still remove idiomatic phrases, and this amounts to a great deal of flattening, particularly in Greek and Latin, where idiom carries much of a passage's nuance and emotional register. Pliny's Letters 9.33 gives a clear example. A word built on the idea of things flowing together, as water flows toward a single point, is rendered in English simply as “flocking,” a term that evokes birds gathering rather than water converging. The metaphor of confluence, of separate streams becoming one current, is lost entirely, and what remains is a serviceable but flattened English word that no longer carries the original image at all, even though it preserves the roughness and directness that a more idiomatic rendering might smooth away. Literal translation is not exempt from this kind of loss merely because it avoids paraphrase; it simply loses a different layer of the original than idiomatic translation does. Both are shadows. They are simply shadows cast at different angles.

Response 3

To the third objection, the extension of the theory of Forms into language gives a precise answer, and this is where the earlier section does its work. The objection assumes that if everything is removed from some prior thing, then no principled distinction can be drawn between different distances of removal, and translation cannot be scrutinised more than anything else. But the removes are not all the same distance. The original text is one remove from the thought that produced it: thought becomes language, once. Memory of that text, or paraphrase of it by someone who has read it, may add a further remove, but it does not have to repeat the entire thought-to-language step twice over in the way translation structurally must. Translation requires the translator to convert the original language back into thought, and then convert that reconstructed thought into a new language. This is not one remove but two, run in the same direction as the very same process that produced the original in the first place. Translation is not merely one more thing that happens to be removed from the truth, alongside memory and paraphrase, in some flat and undifferentiated list. It is specifically and structurally further removed than the original, in the same precise sense that Plato's painter is further removed from the Form than the carpenter is, because the painter copies the carpenter's bed rather than working from the Form directly. Translation copies the original language rather than working from the original thought directly, since the translator has no access to the thought except through the language that already stands between them and it.

Response 4

To the fourth objection: turning something into a shadow can also be an active process, and Plato's own text supports this reading rather than undermining it. The people carrying the statues in the allegory choose to walk along the path with them; someone has built the fire, and someone has decided where the wall stands and how high it is, so that only the upper portion of each statue is visible to the prisoners at all. The shadow is the product of a whole series of deliberate acts, not an accident of light that occurs without any human intervention. Active and passive choices do not, on their own, determine which side of the allegory something falls on. What determines it is whether context is preserved or stripped away in the process, and by that measure, an active and careful process of translation can produce a shadow just as easily as a passive and careless one can. Care in execution is not the same thing as fidelity to the source; a translator can work with great skill and attention and still produce something that is, structurally, a shadow rather than a statue.

Conclusion

The allegory survives its objections because what matters to it is not whether translation is active or passive, singular or plural, but whether the context of the original, and the thought that produced it, is stripped away in the process of rendering it into another language. Literal translation, like the shadow on the cave wall, retains a recognisable outline of its source while losing the surrounding context that gave that source its full meaning, and it does so by structural necessity, since it must pass through the thought-to-language transition twice where the original passed through it only once. Students who are taught only the literal translation, and never shown the original or the interpretive choices standing behind the translation, are in the position of the prisoners described in Book VII: naming what they are given as though it were the whole truth, because it is the only thing they have ever been shown, and because nothing in their education has ever suggested that a fuller, more solid version of the same thing exists just behind the wall.