Based on the linked sources and original analysis. No first-hand playtesting is claimed.
Hint levels
01What the community has proposed
The community guide Language of He Who Watches, written by fartson, interprets square symbols with pairs of gaps as letters, intact squares as spaces, and nested characters as text read from the outside inward and then down columns. It also connects repeated R, L, and X glyphs with right, left, and interact inputs in some wormhole puzzles. These are attributed community findings, not a decoding independently tested for this article. The original chart and final sequences are not reproduced here.
That short description is already a substantial spoiler. If you prefer to discover the writing system yourself, the smaller nudge is simply to record repeated marks accurately and compare where they occur. You do not need a complete alphabet to notice structure. A recurring shape can be useful evidence before you know whether it represents a letter, a separator, or something else.
The rest of this page offers an original workflow for making reliable notes and checking a proposed reading. It deliberately stops short of translating every inscription or exposing the late-game destination described by the guide. A method for reading a clue should let you choose how far to follow it, rather than force every related secret into the same answer.
02Keep three records, not one
Use one record for the symbols as seen, another for a tentative transcription, and a third for the meaning you think the transcription has. These layers answer different questions. The first asks what is drawn. The second asks how it might be represented in ordinary text. The third asks what that text could mean in context. A mistake in one layer should not quietly rewrite the others.
For example, an uncertain gap should remain uncertain in the original drawing even if one possible letter produces a familiar word. Familiar words are persuasive, but they can encourage you to see a mark that is not there. Keep the proposed letter in the transcription layer with a question mark. Then return to the image and check the gap from a clearer view.
This separation also makes correction easier. If a later inscription changes your interpretation of one symbol, you can update the transcription without losing where the symbol appeared. If the meaning of a sentence changes while its letters stay the same, you can revise only the interpretation. Your notes become a record of evidence rather than a single answer that must be defended.
03Choose a clear sample first
Begin with a short, well-lit inscription whose shapes you can distinguish. A long ornate panel may look more important, but it creates more opportunities for copying errors. A modest sample is easier to transcribe twice and compare. The best first sample is one you can inspect again without a long route or a difficult setup.
Give the sample a local label, such as panel beside the entrance, and record its location. Do not use a name that assumes a translation you have not established. Calling it the warning panel can bias later interpretation if it turns out to serve another purpose. A location label remains useful no matter what the text eventually says.
If perspective makes the squares look skewed, move to a clearer viewpoint where you can. Record any part that remains obscured. Do not straighten the entire inscription from memory after leaving. A clean sample reduces uncertainty at the start, which is far easier than trying to diagnose several doubtful characters after a long translation has already been built around them.
Open the guide · contains spoilers
Copy the gaps as carefully as the lines
When a writing system distinguishes similar shapes by missing segments, empty space carries information. Draw the openings explicitly. Note whether a gap lies on a side, near a corner, or at a place that is hard to see. A square that looks intact from a distance may need a closer inspection before you classify it as a separator.
Avoid compressing every shape into a tiny scribble. Leave enough space to preserve the distinction you are studying. If you use a screenshot, make a separate note identifying the relevant boundaries instead of assuming the image alone will remind you. The act of transcription is part of the analysis because it forces you to decide what you can actually distinguish.
Copy the sample a second time without looking at your first drawing, then compare the two. Any disagreement identifies a part worth inspecting again. This is a simple error check, not a claim about how the game's script works. It catches mistakes before they propagate into a proposed alphabet or a long chain of guessed words.
Establish reading order before chasing words
A set of correctly identified symbols can still yield nonsense if you read them in the wrong order. Treat order as its own hypothesis. Record where you begin, how you move through a nested group, and how you move between groups. Then apply the same convention to the whole sample instead of changing it whenever a word becomes inconvenient.
The community guide's outside-in and column-based reading is a useful reported starting point. Keep its application explicit in your notes. If one panel seems to resist the convention, first check whether you copied the nesting and grouping correctly. Do not immediately invent an exception. A transcription error is often easier to rule out than an entirely new grammar.
You can write an ordered list of temporary symbol labels before substituting letters. That lets you test the sequence separately from the alphabet. If the order changes, you can see exactly which group moved. If the order stays stable while the letters change, the unresolved problem lies elsewhere. This division keeps several kinds of uncertainty from collapsing into one vague sense that the translation is wrong.
Build a personal key with confidence marks
Give each distinct observed shape a temporary identifier. Add the proposed letter only when you have a reason for it, and record that reason beside the entry. A letter copied from the community chart has a different basis from a letter you inferred independently across several inscriptions. Both can be useful, but they should not look identical in your notes.
Use a small confidence scale that you can apply consistently. Confirmed by clear repeated observations, plausible but uncertain, and unresolved are enough. Do not assign numerical precision that your evidence cannot support. The purpose is to identify which entries deserve another check, not to pretend that a puzzle notebook is a statistical model.
When two similar shapes seem to map to the same letter, check the drawings before merging them. They may be the same shape seen from different angles, or they may contain a distinction you missed. Keep both entries until the evidence resolves the question. Premature merging can hide the very feature that makes the writing system readable.
Use repetition before interpretation
Look for the same sequence in more than one place. Repetition gives you a cross-check that does not depend entirely on meaning. If you have transcribed the same visible group in two different ways, at least one transcription needs review. That is useful even when you cannot yet read either version.
Repeated groups can also reveal boundaries. A pattern that recurs with a consistent gap before and after it may be a meaningful unit. Record the pattern without immediately deciding what kind of unit it is. It could be a word, a label, or a structural element. The surrounding context will help distinguish those possibilities as your sample grows.
Do not assume that every repeated mark has a linguistic role. Games can reuse visual motifs for decoration or interaction as well as writing. Compare how the mark behaves in context. A symbol embedded in an inscription and a similar shape on an interactive object may be related, but resemblance alone does not establish the exact relationship.
Test a reading against a second sample
Once one sample produces a plausible reading, apply the same rules to a different inscription before treating the system as settled. The second sample should be clear enough that its failure would be informative. If both samples require unrelated exceptions, the first success may have been a coincidence or a product of flexible guessing.
Write down the rules you are carrying over. They might include the order of nested groups, the treatment of a separator, and the current letter mapping. Do not adjust all three at once when the new sample becomes difficult. Change one assumption, see what it affects, and keep a record of whether the change improves both samples or only one.
A strong reading survives contact with material you did not use to invent it. That is the practical value of the second sample. You are not trying to prove an abstract theory of the whole game. You are checking that the method remains useful when it meets a fresh piece of evidence.
Keep ordinary language expectations under control
A partial phrase often invites you to fill in the rest. That instinct can help generate a hypothesis, but it should not replace reading the symbols. Write the guessed completion in a different place from the transcription. Then inspect whether every predicted letter is supported. If the phrase works only by ignoring an awkward mark, keep it tentative.
The same caution applies to grammar. A strange sentence may indicate a copying error, a wrong reading order, or an interpretation you have not understood. It does not automatically mean the game uses a special grammatical rule. Start with the most direct checks before inventing an elaborate explanation for every irregularity.
Context is most useful after the visible evidence is stable. A line near a door may suggest an instruction, while a panel in a quiet space may invite another reading. These are contextual clues, not guarantees. Let them help evaluate competing interpretations rather than determine the transcription in advance.
Separate text from an action sequence
Reading a group of letters and deciding how to act on them are separate steps. The community guide reports an input interpretation for some repeated R, L, and X glyphs. That is a specific reported use, not permission to treat every inscription as a control sequence. First decide whether the local context supports an action at all.
If you do test an action interpretation, record your starting state and the exact input method you used. A letter that you interpret as interact does not by itself establish the default key or button on your setup. Consult your actual bindings and on-screen prompts. This avoids turning a correct symbolic reading into a failed test because the physical input was assumed rather than checked.
Do not continue entering long guessed sequences after an unclear result. Return to the reading and identify which part was uncertain. A failed input can reflect the wrong sequence, wrong starting state, or wrong context. Without separating those possibilities, repeated attempts provide little new information and may expose you to more spoilers than you wanted.
Make corrections traceable
When you revise a character, write why the revision happened. A clearer view showed a second gap is a concrete reason. The new letter made the sentence sound better is a weaker reason that needs another check. Keeping the reason visible helps you avoid treating a convenient guess as an observation.
Review other samples that use the revised character. A local correction may improve several readings, or it may reveal that you combined two different shapes earlier. Either result is useful. Do not stop after fixing the one phrase that drew your attention; the value of a key is that it behaves consistently across your notes.
Preserve an earlier reading if it has already shaped your route or another puzzle theory. You may need to revisit those downstream assumptions. A corrected word can change what you thought an instruction referred to. Trace that consequence explicitly instead of leaving an old plan attached to a new translation that no longer supports it.
Choose your spoiler boundary deliberately
The original community guide discusses material beyond a basic alphabet. Opening it may expose translations and late-game connections you have not reached. Decide what you want before following the link. You might need only one uncertain symbol, a check of the reading order, or a full explanation of a particular inscription.
If another player is helping, ask them to answer only that narrow question. They can confirm whether two copied shapes are distinct without translating the surrounding sentence. They can also tell you whether your reading order is consistent without naming the later location associated with the text. A bounded request often preserves more discovery than searching for a broad complete solution.
When sharing your own findings, separate the method from the result. A post about how you transcribed nested symbols need not place the final translation in its title or opening image. Mark secret material clearly and give readers a chance to stop before the meaning of a late-game panel is revealed.
Handle unresolved inscriptions honestly
Not every public question currently has an answer. The Garden of Motion discussion, for example, records a player's uncertainty about an interior puzzle without a verified solution in the replies at the time of this check. It should not be used to confirm a speculative translation or a particular arrangement of objects.
If your inscription remains unresolved, document the sample and the precise point where the method fails. Perhaps the sequence is clear but the meaning is not. Perhaps one nested group is difficult to distinguish. Those are different problems and deserve different requests for help. An honest unresolved entry is more useful than a confident translation built from missing evidence.
You can return later with another sample or a clearer observation. A partial key and a reliable transcription preserve your progress. The goal is to make each new clue easier to compare, while retaining the freedom to revise the interpretation when the game's evidence points elsewhere.
Audit an ambiguous character in isolation
When a single character blocks a reading, isolate that character before revising the whole sentence. Copy the outer boundary, mark each visible opening, and note the viewing angle. Then find another occurrence that appears to use the same shape. Compare the drawings without looking at the words you hope they form. This reduces the pressure to select whichever letter makes the current sentence easiest to read.
If the two occurrences differ, decide whether the difference is clearly visible or caused by an uncertain view. A confirmed difference deserves a separate entry in your key. An uncertain difference deserves a note and a better sample. Neither result requires inventing a new grammatical rule. You are still working at the level of what the writing actually looks like.
After resolving the shape, return to the ordered transcription and check the affected samples. A stronger character reading may improve more than one inscription, which is useful corroboration. If it improves one and breaks several others, review whether those samples contain the same shape or only a similar one. The broader pattern is more informative than one attractive phrase.
You can ask another player to perform this visual comparison without asking for a translation. Show only the uncertain shape and a second sample, explain which feature is unclear, and request a check of whether they match. This keeps the help focused on observation and can preserve the meaning of the inscription for you to discover.
Finally, keep the corrected character in your key with its evidence. A small note such as checked against two clear samples is enough. You do not need to retain every failed guess, but you should retain the distinction between a shape you have observed confidently and a letter assignment you are still borrowing from the community guide. Those are separate sources of confidence, and keeping them separate makes later corrections easier.