Item #: SCP-7891
Object Class: Euclid
Special Containment Procedures
SCP-7891 is a stone archway located approximately 4.3 km north-east of [REDACTED], Dumfries and Galloway, Scotland. The surrounding 2-hectare area has been designated as a restricted ecological survey site by Foundation front company [REDACTED] Ltd, and is enclosed by a perimeter fence with motion-sensitive lighting directed outward from both approaches.
Access requires Level 2 clearance and written authorisation from the Site-71 research lead. No personnel are to pass through SCP-7891 from east to west during the active period — defined as the 72-hour window centred on each full moon — without full audio-visual monitoring equipment, baseline cognitive testing, and at least one observer remaining outside the anomaly. Personnel who have passed through SCP-7891 during the active period are to complete a 48-hour post-exposure observation period before returning to standard duties.
Post-exposure observation is to include orientation testing, temporal-sequencing interviews, neurological examination, language-output assessment, and preliminary screening for atypical prion markers. Follow-up assessment is required at 30 days, 90 days, and 6 months post-exposure. Subjects presenting with memory discontinuity, inability to produce intelligible English, spontaneous Gaelic-language competence, or suspected neurodegenerative symptoms are to remain under Site-71 Medical supervision until cleared by the research lead and Medical.
Crossing the third arch — oriented north-west, bearing the unscribed face — is prohibited under all conditions regardless of lunar phase. Personnel who cross the third arch during the active period are considered temporally displaced until confirmed otherwise. A standing recovery protocol is in place at Site-71. Personnel who cross the third arch outside the active period have, to date, reported no anomalous effects. This is not confirmation that no effects occur.
Object class review has been requested. It has not yet been completed. Euclid classification is maintained in the interim.
Subjects who report being addressed during a crossing, per the phenomenon documented under Description, are to be flagged for additional psychological follow-up. Re-entry requests that cite the address as a motivating factor are to be escalated to the Site-71 research lead rather than processed under standard review.
Subjects affected by the language-output presentation of Post-Liminal Neurological Syndrome (PLNS-L) are not to be compelled to communicate in English except during approved diagnostic sessions. External communication with PLNS-L subjects is to be conducted in Scottish Gaelic by certified personnel or through vetted translation software. Affected subjects consistently report thinking in English and understanding English normally; however, attempts to speak, write, type, or sign English produce severe dyslexic and dysgraphic output to the point of functional gibberish. Gaelic output remains coherent. This is not to be interpreted as non-compliance.
A recording station is to be maintained at the site at all times during the active period. Logs are to be cross-referenced against the lunar calendar monthly and submitted to the Site-71 archive.
SCP-7891 is not to be structurally disturbed. Attempts to relocate, disassemble, or otherwise alter the archway are on hold pending further understanding of the anomaly's mechanism.
Medical personnel should note that suspected PLNS presentations are low-frequency and delayed. A normal 48-hour observation period does not rule out later manifestation. No confirmed secondary transmission has occurred through ordinary contact, spoken Gaelic, written Gaelic, or exposure to affected English-language output. Standard prion-containment precautions remain mandatory when handling biological samples from PLNS-CJD cases.
Description
SCP-7891 is a free-standing stone structure in a moorland clearing in Dumfries and Galloway, Scotland. It is composed of three continuous granite arches, each a complete curve, intersecting the other two — the stone of each arch passing through the physical mass of the others at two points per pairing. The structure stands approximately 2.1 metres at its apex. No mortar is present. No joinery is visible at the intersection points. The structure should not be standing. It is.
The three arches are oriented at equal angular intervals, producing six individual threshold crossings and three zones where pairs of arches overlap — spaces that are simultaneously inside two arches at once. Foundation engineers noted the geometry is consistent across all axes of measurement and inconsistent with load-bearing principles on record.
Thermoluminescence testing places the most recent significant heat exposure between 1680 and 1710 CE. The structure shows no signs of having been assembled from separate pieces.
A minority of exposed subjects have developed delayed neurological sequelae now provisionally grouped under Site-71 Medical designation Post-Liminal Neurological Syndrome (PLNS). Recorded presentations include short-duration temporal amnesia, progressive prion pathology clinically resembling a mutated form of Creutzfeldt-Jakob disease, and selective failure of external English-language production. The latter presentation is linguistically paradoxical: affected subjects retain English-language thought and comprehension but can only communicate coherently in Scottish Gaelic. English speech and writing are produced as dyslexic, dysgraphic, or phonologically disordered strings that cannot be reliably reconstructed even by the subject who produced them.
The Inscriptions
All inscriptions face inward, visible only in the act of crossing.
On the underside of the shared apex: 'S e fìrinn liminalachd. The Foundation's consulting linguist renders this as "liminality is the truth" or more precisely "the truth is liminality," the copular construction placing emphasis on the predicate. Liminalachd — formed from the Latin limen via the Gaelic abstract-noun suffix -achd — does not appear in any documented period source. The consulting linguist's note reads: "Either the word is older than the record, or the record is wrong about when the word was carved. I cannot tell you which."
Following identification of PLNS-L, the same linguist appended a second note: "The structure does not merely prefer Gaelic. In the affected cases it appears to leave English available internally and remove it at the point of exit. That is not language acquisition. That is a door closing in one direction."
On the inner face of the first arch — oriented east to west — a triskelion: three spiral arms, each completing a full revolution, clockwise. The stone around it has been deliberately smoothed.
On the inner face of the second arch — oriented south-east — a Triple Goddess symbol: waxing crescent, full circle, waning crescent in a single horizontal line, the full circle marginally larger. The same deliberate preparation of the surrounding stone.
The inner face of the third arch — oriented north-west — bears no inscription. The stone has been smoothed identically to the other two.
The Address
Every subject who completes a crossing, through any of the three arches, reports being addressed at the moment of completion. This holds across all 23 first-arch subjects, all second-arch subjects, and all third-arch subjects, including those whose third-arch crossings produced encounters with other people.
The address is brief. Subjects describe a single phrase, or occasionally two, spoken in a tone consistently characterised as approving. The specific wording is reported inconsistently — "good," "well met," "you have done well," "welcome back" — but the consistency is in the register rather than the words: every subject describes being told, in effect, that they have done something correctly, or that their having crossed is itself a thing to be commended.
Eleven subjects report the address in English. Six report it in a form of Gaelic the consulting linguist has been unable to date precisely, but does not regard as contemporary. Four report no specific language at all — a meaning received without words attached to it, in the manner of understanding a tone of voice from another room before making out what is said. Two declined to specify.
These figures refer to immediate post-crossing self-report and should not be confused with later PLNS-L presentation. In three cases, subjects who initially reported the address in English became unable to reproduce that account in English within 11 days. All three retained the content of the memory, stated that they continued to think about it in English, and then gave coherent accounts only when permitted to speak Gaelic. English reproductions from these sessions are included in the medical archive but are not transcribed here, as they consist primarily of unrecoverable orthographic distortion and phonetic collapse.
No subject has identified the source. No subject reports seeing anyone speak. Several have described the voice as coming from no particular direction, or from every direction at once, which is consistent with reported sensory changes during the threshold state, but has also been documented by second- and third-arch subjects, who do not otherwise report sensory alteration.
The address has been picked up on recording equipment in 9 of 23 first-arch instances and in 2 of the documented third-arch instances. In each case the recovered audio is faint, brief, and resistant to spectral analysis — the Foundation's audio team has been unable to isolate a vocal source, a direction of origin, or, in 6 of the 11 recordings, a confident transcription. The recordings exist. The Foundation does not currently know what is on them.
No subject has reported the address as unwelcome. Three subjects, all from the first arch, have described it as the most significant part of the crossing — more significant, in their accounts, than the morphological change itself. A fourth subject, later diagnosed with PLNS-L, wrote the phrase math a rinn thu 28 times during observation. When asked to translate the phrase into English, the subject wrote eight lines of non-lexical text and then became distressed, stating in Gaelic that the English version was "clear in my head and dead in my mouth."
This phenomenon is documented separately from the threshold state, the lunar synchronisation, and the temporal displacement, because it does not vary meaningfully between them. The three arches differ in what they show a subject. They appear not to differ in this.
The First Arch — The Threshold State
Subjects who cross the first arch during the active period — the 72-hour window centred on each full moon — enter the threshold state.
The threshold state is lycanthropic in character. The Foundation has avoided this word in previous documentation. It is the correct word.
Observers and recording equipment consistently document partial morphological transformation in affected subjects: elongated jawlines, altered pupillary shape, skeletal restructuring of the extremities, changes in hair distribution. The degree of change varies. None of it is cosmetic. In three cases, subjects have demonstrated sensory capabilities during the threshold state — acuity of hearing, sensitivity to scent — that significantly exceed human baseline and have not fully resolved at the 72-hour mark.
First-arch exposure is also associated with two documented PLNS presentations, though causation has not been established. Subject-06-M developed a 17-minute temporal amnesia covering the resolution of the threshold state. Subject-21-T developed PLNS-L nine days after exposure, retaining English comprehension and internal English thought while losing all coherent English output. Subject-21-T had no recorded Gaelic proficiency before exposure and now refuses English-language diagnostic prompts on the grounds that "the words arrive broken."
Subjects themselves report no pain. What they report instead is presence — a quality of simultaneous existence in two states that none have described as distressing in the expected sense. Seventeen of 23 subjects have used the word "both." Eight found the experience distressing. Fifteen did not. Three have formally requested re-entry.
The threshold state is not transformation. It is not lycanthropy in the folkloric or clinical sense — there is no loss of self, no dissociation, no violence. It is something older and more specific than the folklore describes, which is: the state in which a person is revealed to be two things at once, held in that revelation long enough to know it, and returned to their singular self at dawn with the knowledge intact.
The phrase "knowledge intact" has been retained from earlier documentation but should be treated cautiously. Affected subjects generally retain selfhood, procedural memory, and autobiographical continuity; however, small discontinuities around dawn-resolution events have been documented. These are temporal rather than dissociative. Subjects do not report becoming someone else. They report that a few minutes or hours are missing from the crossing's edge, as though the return was edited after the fact.
The Foundation has classified subjects who show more pronounced morphological change, or whose sensory alterations persist longer post-exposure, as provisionally designated persons of interest. Seventeen of 23 subjects fall into this category. The remaining six showed minimal change and do not describe the threshold state in the same terms. The distinction has not been formally investigated. It should be.
The Second Arch — Lunar Sensitivity
Subjects who cross the second arch during the active period report a qualitatively different experience from the first arch. Morphological changes are not observed. What subjects consistently describe is perceptual synchronisation with the lunar phase at the moment of crossing.
At full moon, subjects describe a state of overwhelming presence and clarity — several have used the word "called." At the waxing phase, subjects describe anticipation without object, a pull forward without destination. At the waning phase, subjects describe a quality of release — not loss, but setting down.
The second arch has no effect outside the active period. Two subjects have crossed it at the new moon, during the active window's outer edge. Both reported the threshold going dark. Neither elaborated. Both requested not to elaborate.
One of these subjects later developed transient PLNS-L lasting approximately 31 hours. During that interval the subject understood all English instructions, followed written English protocols correctly, and completed non-verbal reasoning tasks without deficit. Spoken and written English output was unintelligible. When addressed in Gaelic by Foundation linguist Dr. M. Ealasaid, the subject answered with fluent Scottish Gaelic and gave the only currently coherent description of the new-moon crossing: "I was not absent. I was where a person is before language leaves." The subject reported thinking this sentence in English before saying it in Gaelic.
No morphological changes have been documented for second-arch crossings. The Foundation has tentatively concluded the second arch is the lesser anomaly. Dr. Gunn has formally disagreed with this conclusion and her disagreement is noted in Addendum 7891-B. Since the addition of language-output assessment to the containment protocol, this conclusion has been placed under review.
The Third Arch — Temporal Displacement
Subjects who cross the third arch during the active period are displaced temporally within a range now established as 693 to 1472 CE, with all documented cases resolving within the geographical boundaries of what is now Scotland. They emerge into the same clearing, in that era. The archway is not present. The clearing is. They return at dawn of the displaced era, or dawn of the present, or both — the question has not been resolved.
The range was not initially understood to be a range. Early documentation assumed the clustering around 1306 to 1314 was the extent of the phenomenon. Subsequent cases have revised this significantly. The full temporal distribution is recorded in Addendum 7891-E. The clustering has not disappeared with the expanded range. It has become more complicated.
No subject has failed to return. One subject returned with a fragment of woven textile dated between 1280 and 1340 CE. One subject returned with a small iron pin of a type produced in the Pictish period, predating 900 CE. One subject returned with mud on her boots that contained pollen from a plant species confirmed extinct since approximately 1100 CE and not present in the current moorland. One subject returned with a burn scar on her left forearm that was not present before crossing and that she says she does not remember receiving. Two third-arch subjects returned with temporal amnesia exceeding six hours. In both cases, subjects remained calm, oriented to person and place, and able to answer questions about the current year. They could not account for intervals that were subsequently confirmed to have occurred during displacement. One subject described the missing period, in Gaelic, as "time that remembered me without bringing me along."
The Foundation has noted these instances. The Foundation has not determined what they indicate about the nature of the displacement or the extent to which subjects are participants in the displaced era rather than observers of it.
Third-arch exposure is the only exposure category currently associated with confirmed PLNS-CJD. Two subjects have shown progressive neurological decline, abnormal cortical ribboning, myoclonus, and prion-seeding activity in cerebrospinal fluid. Sequencing has not matched any known human prion disease. Site-71 Medical uses PLNS-CJD as a provisional designation because the presentation clinically resembles Creutzfeldt-Jakob disease, not because the pathology is understood. Standard decontamination procedures have prevented secondary cases.
Crossing the third arch is prohibited under all conditions regardless of lunar phase.
Historical Context
During a standard archive review sweep in 2019, Foundation archivists recovered three documentary sources from the Church of Scotland archive in Edinburgh relating to the archway's location.
The first is a parish record from 1689 describing a formal ecclesiastical trial conducted at the site, at which three individuals — named as Alasdair Mhic Aoidh, Mairead Nic Iain, and Eòghann Domhnallach — were charged with lycanthropy. The record states that all three were "discovered in states of manifest bestial transformation at the limn, upon the night of the full moon." The outcome of the trial is not recorded in the surviving document. The archway is referred to throughout as an cusp — the threshold — and in one instance as an limn, from the Latin limen. It is unclear from the record whether this was the presiding minister's coinage or a name the location already carried.
A damaged marginal note in the same hand as the trial record describes one accused person as "answering sensibly in the Irish tongue while all English came out crooked and no wise legible." Foundation palaeographers initially treated this as a reference to poor literacy or deliberate evasion. After identification of PLNS-L, the note has been reclassified as a probable historical symptom record.
The second source is reproduced in Addendum 7891-A.
The third is a letter from a Church of Scotland minister to a colleague, dated 1691, two years after the trial. It reads, in its entirety: The stone stands. Thee ought not to attend it by night. Both correspondents are identifiable by name. Neither appears in subsequent parish records.
Addendum 7891-A — Recovered Document
The following document was found folded inside the back cover of the 1689 parish record. It is written in English throughout, though several Gaelic words appear untranslated. The handwriting is consistent with the late seventeenth century. The author is unknown.
I have been through the arch three times now. The first time I did not understand what was happening to me. The second time I did not believe it. The third time I went deliberately, under Her full face, because I had a question I could not answer from the outside.
The question is this: which one is true?
There is the shape they know. The name my father gave me, the face that fits it, the fire tended, the anxious waking before dawn. A person by any reasonable account. And then there is the other thing — which is not a replacement for the first, as I had always assumed when I allowed myself to think about it at all, but something that was already there underneath the surface of the skin. Present. Patient. Older than the name.
The arch does not create the second thing. I understand that now. It only holds you in the moment of knowing it, long enough that you cannot look away.
The people who tried us at the limn called it possession. They said something had entered us from without. They were wrong — completely, demonstrably wrong — and they did what they did anyway, which is the particular cruelty of people who have already settled on the truth before examining the thing itself.
Nothing entered me at the arch. I did not gain what I did not already carry. What I lost was the ability to pretend I was only one thing when I have always been two.
Mairead says we should leave this valley. She is almost certainly right. I find I am not yet done with the question.
Which is true — the man at this table, writing this, or what I become under Her full light?
I have been thinking the answer might be: neither. I have been thinking the answer might be the arch itself. The crossing. The space between one step and the next, where both feet are briefly off the ground and you are committed to neither side.
I have spent the whole of my life being afraid of that space. I am trying to find a way not to be.
The arch does not give you the wolf. The wolf was there. The arch gives you the moment of knowing it, held open long enough to step into rather than away from. That is either a gift or a test and I have stopped being certain there is a difference.
Selene keeps the night. I trust Her with it.
'S e fìrinn liminalachd.
Addendum 7891-B — Researcher Notes, Dr. A. Gunn, Site-71
Filed 3 September 2023, with amendments 14 November 2023 and following.
I want to raise something in the recent subject self-reports that the summary statistics are not capturing adequately.
Across all 23 documented cases, the language used to describe the threshold state has been remarkably consistent. "Both." "Between." "At the same time." "Neither one thing nor the other, but somehow both at once." These are not prompted responses — debriefs use open-ended questions with no suggested vocabulary. Twenty-three subjects, different ages, different backgrounds, no common prior connection, all arriving independently at the same description. That is the kind of consistency that usually means something.
The variation is in whether they found it distressing.
In 8 of the 23 cases, subjects reported distress during the threshold state. In the remaining 15, they did not. Several in that second group used the word "relief." One said it felt like "seeing myself in a mirror that was telling the truth for once." Three have submitted formal written requests to re-enter the archway.
I am not proposing a reclassification — Euclid remains correct and the morphological data alone justifies it. What I am raising is that the Foundation's current harm-assessment framework asks us to identify damage and stop there, and SCP-7891 is producing a consistent secondary effect that does not fit into that category. Fifteen of 23 subjects entered a documented anomalous event and came out feeling better than when they went in. That is a data point. The framework should account for it.
The three re-entry requests have not been accommodated, which is the right decision for now. I note that two of those three subjects submitted in writing, twice.
There is another pattern Medical keeps insisting is adjacent rather than central. I am not convinced.
We now have subjects who understand English perfectly and say, through interpreters, that they continue to think in English. They read English instructions, solve English word problems internally, and respond correctly to non-verbal tests based on English prompts. Then they try to speak English and what comes out is not English. Not broken English. Not aphasia in the ordinary sense. A chain of sounds that look, on transcription, as if the concept of spelling has been put through a mill. Written English is worse. Catastrophic dyslexia, catastrophic dysgraphia, but only at the point where thought leaves the skull.
Then they are asked in Gaelic, and they answer. Fluently. Calmly. Sometimes more precisely than they ever spoke English before. Two of them had no Gaelic beyond road signs and school songs. One had no recorded exposure at all.
I asked Subject-21-T what language she was thinking in. She answered in Gaelic: English. I asked why she was not speaking English. She answered: because English breaks when I open my mouth.
That should not be a footnote. That is the same structure as the archway. Something remains true internally and changes at the threshold.
The archway has been standing since at least 1689. Whatever it is doing, it has been doing it for a long time. The harm it has caused has not, as far as I can determine, been located where we first expected to find it.
I am less comfortable saying that after reading the first PLNS-CJD file. The disease is real. The deterioration is real. I am not trying to turn an anomalous prion disorder into poetry because I like the word liminal. But I also do not think Medical understands why a clearing associated for eight centuries with doubled bodies, lunar perception, lost time, and impossible Gaelic has now produced a prion conformation that cannot decide whether it is the patient's protein or a historical artefact. I know that is not how proteins work. I have read the report. The report also does not know how the protein works.
The parish record describes three people tried at the limn. The outcome is not in the surviving document. The letter says the stone still stands and advises against attending it by night. Neither of the correspondents appears in subsequent records.
The triskelion and the Triple Goddess symbol have been in the infrared survey data since initial documentation. Both face inward. The anomaly activates at full moon. I would like to formally put on record what the description section declines to say plainly: the archway was marked by someone who understood what it does. That is not a comfortable conclusion. It is, however, the most parsimonious one currently available, and I think the file should say it.
I wrote previously that the third arch was blank in the way that the new moon is dark — not yet visible, something coming. I was wrong about the direction. The third arch does not point forward. It points back to a period three and a half centuries before the archway was built, to a clearing where the archway does not yet exist and something was already happening.
The triskelion turns. The moon moves through its phases. The third face shows what cannot be inscribed because it predates the inscription. The blank is not a gap. It is the part of the cycle that this structure did not build — the part that was already there when the builders arrived.
I would like to formally revise my earlier framing. I would also like to note that this makes the 1689 trial significantly more complicated than the parish record suggests. Three people were charged with lycanthropy at the limn. The parish record is focused on the threshold state. I no longer think that was the whole of what the Church of Scotland witnessed.
I do not think SCP-7891 is dangerous in the way we initially assumed it might be. I think it is dangerous in the way that accurate mirrors are dangerous — which is a different category, and one the Foundation does not currently have good tools for.
I continue to recommend Euclid classification.
— Dr. A. Gunn, Research Lead, Site-71
Amendment, 14 November 2023:
The original note stands. I want to add to it.
I want to address the description section's characterisation of the second arch as the lesser anomaly, because I think it is wrong in a way that matters.
The first arch shows you what you are. The morphological changes, the sensory alterations, the doubling — these are the arch externalising something internal. The seventeen subjects who showed significant change were not changed by the arch. The arch found something already present in them and held it up to the light long enough to be seen. The six who showed minimal change — I do not think the arch found nothing. I think it found something that does not express in those terms.
The second arch synchronises you with the moon. The Foundation has read this as the lesser effect because it produces no visible morphological data. I think the Foundation is measuring the wrong thing. The subjects who crossed at the waning phase and described setting something down — two of those subjects were among the three who formally requested re-entry to the first arch. They did not request re-entry to the second arch. They said they were finished with it.
The two subjects who crossed at the new moon and declined to elaborate — I have spoken informally with both of them, outside the debrief structure. Neither will discuss what they experienced. Both are, by any observable measure, fine. One said to me: "I wasn't there. That's not a description of what it felt like. That's what happened. I wasn't there."
I do not know what to do with that. I am putting it in the record because the record should have it.
The third arch I have already addressed. I was wrong about the blank face pointing forward. It points back. The moon moves through phases. The triskelion turns. The third arch opens onto a time when none of this existed yet and something was in that clearing anyway. That is not a lesser effect. That is not a greater effect. It is a different axis entirely and the current framework does not have a word for it.
The word in the keystone is liminalachd. The truth is liminality. I keep returning to the copular construction — not "liminality is true" but "the truth is liminality," as though the predicate is the stable thing and everything else is what changes around it. Someone built this structure and inscribed that sentence on the part that all three arches share, and then carved the moon's phases on one arch and a symbol of cyclical return on another and left the third face empty, and the empty one is the one that opens backward in time.
I think the structure is a single argument made in three registers. I think we have been documenting its components as though they were separate anomalies. I do not think they are separate anomalies.
I still recommend Euclid classification. I recommend it less confidently than before.
— Dr. A. Gunn
Final note, appended following completion of Addendum 7891-E:
I have now spent four months compiling that record and I want to say something plainly before I close the file on it.
The clearing predates the archway by at least a millennium. Possibly longer — the Pictish stone is the oldest direct reference and it is not necessarily the first. The arch was built in a clearing that was already known, already named, already treated with a specific combination of reverence and avoidance across eight centuries of continuous documentation. The wolf-stones. The wolf-clearing. The three-stone moor-place where the wolf-natured gather.
Whoever built the archway in the late seventeenth century did not create the anomaly. They formalised it. They inscribed the keystone with a word that shouldn't exist yet and carved the moon's phases on one face and a symbol of cyclical return on another and left the third face blank, and what they built was an answer to something the clearing had been asking for a very long time.
I do not know what the clearing was before the arch gave it a frame. I know it was already doing something. The Pictish stone shows the transformation. The Old Irish gloss calls it the thing that wears a wolf. Eight centuries of documentation says the same thing in different registers: people who came here under the full moon came back different, and the difference was not removal but addition, and nobody — not the Pictish carver, not the Old Irish annotator, not the Benedictine herbalist, not the bardic poet — described it as harm.
Then three people were tried here in 1689 and the record doesn't say what happened to them.
I think we are documenting the wrong question. I think the question is not what SCP-7891 does but what it has always been for, and I think the answer to that is written in a language we keep translating as anomaly when the word it actually uses is truth.
'S e fìrinn liminalachd.
I still recommend Euclid classification. The classification is accurate. It just isn't sufficient.
— Dr. A. Gunn
Further note:
I have called this structure dangerous in the way an accurate mirror is dangerous. I need to revise that once more.
An accurate mirror does not speak. This one does. Every subject who completes a crossing — all three arches, no exceptions in the data so far — is addressed, briefly, in terms of approval. Not instruction. Not warning. Praise, more or less, for having crossed at all.
I do not know what to do with a mirror that holds you to the truth of yourself and then tells you that you have passed. I am not assuming benevolence. I am noting that the structure has had eight centuries and three documented governing bodies' worth of opportunity to be unkind to the people who come to it, and has, as far as the record shows, never once taken it.
I am also noting, because the file needs the complication, that some subjects come back damaged. Not punished, I think. Damaged. Missing hours. Unable to let English cross the boundary between thought and voice. Carrying proteins in their nervous systems that behave like bad translations of themselves. If the arch is a mirror, the mirror has edges. People can cut themselves on accurate things.
— Dr. A. Gunn
Addendum 7891-C — Interview Log
Subject-19-H, 48-hour post-exposure debrief. Interviewer: Dr. A. Gunn.
Gunn: Walk me through what you experienced. From the moment you stepped through.
Subject-19-H: It happened immediately. Not gradually. One step I was looking at the archway, the next I was — both things.
Gunn: Can you be more specific?
Subject-19-H: I could hear everything. The wind through the heather from half a kilometre away, a bird I couldn't see. But I was still me. Still thinking in sentences, still knowing my name. Nothing was missing. It was more like — [pause] — like something had been added back that I hadn't known was gone.
Gunn: Were you aware of the morphological changes observers documented?
Subject-19-H: No. I thought I looked normal. [pause] Do I still not look normal?
Gunn: You're back to baseline. When did the threshold state resolve?
Subject-19-H: At dawn. I was sitting by the perimeter fence and I could feel it — not painfully, just — like a tide going out. I knew which one I was again.
Gunn: How did that feel?
Subject-19-H: [pause] Like putting one shoe back on.
Gunn: What happened between dawn and when the perimeter team found you?
Subject-19-H: I was at the fence.
Gunn: That is where they found you, yes. There are eleven minutes on the exterior recording where you are visible but not responsive.
Subject-19-H: [pause] I don't remember that.
Gunn: Do you remember hearing anything during that period?
Subject-19-H: No. I remember the tide going out. Then I remember being cold. If there was time between those, I don't have it.
Gunn: Does that frighten you?
Subject-19-H: Less than it should.
Researcher note: Subject-19-H submitted a re-entry request twelve days after this interview. Request is pending standard review. The eleven-minute discontinuity was initially treated as ordinary post-anomalous disorientation. Following later PLNS case review, it has been reclassified as probable transient temporal amnesia.
Addendum 7891-D — Interview Log
Subject-23-F, 48-hour post-exposure debrief following third-arch crossing. Interviewer: Dr. A. Gunn.
Gunn: What did you see when you came through?
Subject-23-F: Trees. The clearing was smaller — younger trees around the edge, or different ones, I couldn't tell. The archway was gone. I turned around and there was nothing there. Just moorland.
Gunn: Were there people?
Subject-23-F: Not immediately. After — I'm not sure how long, the light was low — a group came through the tree line from the south. Six or seven of them. Armed. They looked at me and then at the clearing and then they kept moving.
Gunn: They didn't approach you?
Subject-23-F: One of them stopped. An older man. He looked at me for quite a long time and then he said something I didn't catch. [pause] Not English. Not quite what I'd call Gaelic either, but closer to that.
Gunn: And then?
Subject-23-F: He moved on with the others. I sat in the clearing until dawn and came back through.
Gunn: You said he looked at you for a long time. How would you characterise the look?
Subject-23-F: Like he'd seen something in that clearing before. Like he wasn't surprised.
Gunn: Can you repeat what he said?
Subject-23-F: I don't know the words. I know what they meant.
Gunn: Try in English.
Subject-23-F: It was — [unintelligible vocalisation removed from transcript] — no, that's not right. That's not — [subject becomes visibly distressed.]
Gunn: Stop. Try the closest version you can manage.
Subject-23-F: [in Gaelic] He said the stones were not there yet, but the place knew me.
Gunn: You speak Gaelic?
Subject-23-F: [in Gaelic] No.
Gunn: What language are you thinking in now?
Subject-23-F: [in Gaelic] English. I can hear it properly inside. It breaks when I send it out.
Researcher note: Subject-23-F produced no coherent English speech for approximately 41 minutes following this exchange. Written English output during the same interval consisted of repeated malformed grapheme clusters, several of which approximated English words but did not retain stable spelling between attempts. Subject remained able to read English instructions and answer them in Gaelic. Full PLNS-L onset was later confirmed.
Addendum 7891-E — Temporal Incident Record
Compiled by Dr. A. Gunn, Site-71. Cross-referenced against Foundation historical database, Scottish national archive records, and ecclesiastical records held at Edinburgh and St Andrews.
What follows is a record of events — documented in period sources, recovered through Foundation archive sweeps, or reported by subjects — that the Site-71 research team now believes relate to SCP-7891 and the clearing at its location. They are presented in chronological order of the displaced era, not in order of discovery.
c. 693 CE — Pictish settlement records, provenance uncertain
A Pictish carved stone recovered from a site approximately 1.8 km from the clearing depicts, among standard Pictish symbols, a figure in apparent mid-transformation — lower body human, upper body indeterminate, arms raised — standing within what the Foundation's art historian consultant describes as "a triple arch or a three-part enclosure, rendered in a style consistent with the period but unlike any documented Pictish architectural subject." The stone is currently in a regional museum under a standard Foundation cover designation. It predates the archway's thermoluminescence range by approximately a millennium.
Subject-23-F, debriefed following the third-arch crossing in which she encountered the armed group, was shown a reproduction of the Pictish stone without context. She identified the triple-arch element immediately. She said: "That's the clearing. That's what it looks like from inside when all three arches would be there, if they were there." She was not asked a follow-up question. The interviewer did not have one.
During the same session, Subject-23-F attempted to write a caption for the figure in English. The resulting line was graphically consistent with severe acquired dysgraphia and could not be transcribed. When instructed in Gaelic, she wrote: Chan eil clach ann fhathast. Tha an t-àite ann. Translation: "There is no stone yet. The place is there."
843 CE — Annals of the Kingdom of Dál Riata, fragment
A marginal notation in a manuscript fragment held at the Foundation's Edinburgh archive — not the main text, the margin — reads, in Old Irish: the wolf-clearing holds its door open at the moon's full face, and those who pass through it pass through themselves. The manuscript's main text concerns the consolidation of the kingdom under Cináed mac Ailpín. The notation is in a different hand from the main text and has no attribution. It is not a gloss on the surrounding passage. It does not relate to anything on the page.
A second marginal mark on the fragment, previously catalogued as scribal damage, has been re-examined following PLNS-L identification. The mark appears to be a failed attempt at Latin glossing in which individual letters are recognisable but their order is unrecoverable. The same hand then writes a coherent Old Irish sentence beside it: the old speech remains where the new speech falls apart. The Foundation's linguist has advised against over-interpreting this line. Dr. Gunn has advised against ignoring it.
The phrase the wolf-clearing in Old Irish — réidhlean an mhadaidh-allaidh — uses a term for wolf that carries a secondary meaning closer to "the thing that wears a wolf" than "wolf" in the straightforward sense. The Foundation's linguist flagged this without elaborating. Her follow-up note reads: "I know how that sounds. It is still the accurate translation."
c. 1050 CE — Possible subject return artifact
The extinct pollen recovered from Subject-17-K's boots has been dated with reasonable confidence to the period 1000 to 1150 CE. The plant species — a variant of Betula not documented in any surviving botanical record — appears in one other source: a monastic herbalist's notes from a Benedictine house in southern Scotland, c. 1100 CE, which describes a plant found near "a clearing of three joined stones on the moor" with properties the herbalist characterises as useful for "those caught between forms." The herbalist does not elaborate on what "between forms" means. The plant is described as growing only within approximately thirty metres of the stones.
The plant is not present in the current clearing. The clearing has been surveyed. It is not there.
1263 CE — Battle of Largs, peripheral record
A Norse account of the aftermath of the Battle of Largs — held in the Foundation's Scandinavian archive — contains a passage describing a group of Scottish fighters encountered in the days following the battle who were "not wholly men in the moonlight" and who "came from the moorland stones." The account is a personal letter, not a military record, and is generally treated by historians as embellishment. The Foundation notes that the letter writer's described route from the coast to his ship passes within four kilometres of the clearing. The letter writer does not describe encountering the stones himself. He describes being told about them by a prisoner.
1314 CE — Bannockburn, outlying reports
Three separate sources from the period immediately following Bannockburn reference the clearing directly, more explicitly than any prior documentation. A chronicle held at St Andrews describes "a place of three arches on the moor of [REDACTED] where men become other than themselves under the full moon and are not harmed by it." A private letter from a nobleman to his wife, recovered from a private collection and now in the Foundation's archive, states that two members of his company crossed "the wolf-stones" on the night of the full moon before the battle and "came back changed in no way that could be named but in every way that could be seen." A third source — a monastic account — describes a woman found in the clearing at dawn, disoriented, wearing clothing "of a strange cut" and unable to account for where she had come from. The account further states that the woman "held the speech of the Gael with sense, yet when urged toward the Saxon tongue made marks that no clerk could keep straight."
The Foundation has cross-referenced the third source against current personnel records. The description of the clothing is consistent with Foundation-issue field gear. The language description is consistent with PLNS-L. The date of the monastic account corresponds to a full moon. No personnel were assigned to SCP-7891 on that date. Clearance for third-arch crossings was, at that time, not yet restricted.
Access logs from the relevant period have been requested from Site-71. They are pending.
1349 CE — Black Death records, Dumfries and Galloway
Parish death records from the Black Death's passage through Dumfries and Galloway note that the settlement nearest the clearing reported significantly lower mortality than surrounding settlements — approximately 30% of the regional average. The parish priest's account attributes this to "the mercy of God and the protection of the old stones on the north moor, which the people say do not permit certain crossings." The Foundation has flagged this for investigation and has not yet investigated it. This has been in the flagged queue for three years.
c. 1390 CE — Possible extended displacement
Subject-09-R, who crossed the third arch in 2021 and returned apparently normally, submitted a supplementary report six months after the event. The report consisted of thirteen handwritten pages. The subject stated that she had written it in English. Forensic linguistics determined that approximately 96% of the text was unrecoverable, exhibiting extreme dysgraphic distortion without stable word boundaries. When asked to provide the report orally, Subject-09-R gave a coherent account in Gaelic. In it, she describes a recurring dream — beginning the night of the crossing and continuing without interruption — in which she is present in the clearing in a period she estimates, from the clothing of the figures she sees, as late fourteenth century. In the dream she is not a visitor. In the dream, she has been there for some time and knows the land around the clearing well.
Subject-09-R was referred for psychological evaluation. The evaluating clinician's report is attached. The relevant line reads: "Subject-09-R presents no symptoms of dissociation, psychosis, or trauma response. She is sleeping normally and functioning normally. She would simply like the dreams to stop. She has also noted, with some discomfort, that she appears to know the names of people she has never met."
The dreams have not stopped. Subject-09-R has been seen twice since. At the most recent session she said, in Gaelic and unprompted: "I think I was there. Not in the dream. I think when I crossed, part of me was there and came back and the rest of me didn't notice." She paused and then added: "I don't think it's the bad part of me that stayed." Subject-09-R continues to report thinking in English. All English output remains functionally gibberish.
1411 CE — Battle of Harlaw
A single line in a Gaelic bardic poem composed in the aftermath of the Battle of Harlaw — a poem otherwise concerned with grief and the cost of the Highland charge — reads, in translation: and the wolf-stones on the south moor remember every one of them. The poem does not mention the wolf-stones before or after this line. The translator's note reads: "Interpolation seems unlikely. The line scans correctly. I don't know what it's doing there."
1472 CE — Final pre-construction record
The latest pre-archway documentary reference to the clearing — the thermoluminescence range placing construction at 1680 to 1710 CE — is a single entry in a burgh record from 1472. It lists, among other local landmarks to be avoided after dark, "the three-stone moor-place where the wolf-natured gather at moon-full." A later copyist's note adds: "and where the English hand writes crooked if the moon has looked back." The burgh record is otherwise administrative. It is a list of where not to walk at night. The clearing is between a river ford and a mill.
The Foundation notes that this is the last historical reference to three stones. All subsequent references, until the 1689 trial record, describe a single archway. Sometime between 1472 and 1689, the structure either changed form or changed what people perceived of its form. These are not the same thing, and the Foundation has not yet determined which one happened.
Addendum 7891-F — Interview Log
Subject-19-H, supplementary interview conducted following formal re-entry request. Interviewer: Dr. A. Gunn.
Gunn: You submitted a re-entry request twelve days after your original debrief. I'd like to understand why.
Subject-19-H: I left something out the first time. I don't think I was hiding it. I don't think I had the words yet.
Gunn: Go on.
Subject-19-H: Right at the end. After the tide-going-out feeling, before I was fully back. Something said something to me.
Gunn: What did it say?
Subject-19-H: "Good." Just that. Maybe "well done," I'm not sure now, it's already started to blur, but it meant good.
Gunn: Do you know who said it?
Subject-19-H: No. I didn't see anyone. I wasn't frightened by it, if that's what you're asking.
Gunn: What were you frightened by, if anything?
Subject-19-H: Nothing in there. I'm a bit frightened now that I want to go back to hear it again.
Gunn: Can you say more about that?
Subject-19-H: [pause] Nobody's said that to me in a long time. Not about anything that mattered. I don't know what it is. I don't think it matters what it is. I'd like to hear it again.
Gunn: I need to ask about the written note attached to your re-entry request. Did you write it yourself?
Subject-19-H: Yes.
Gunn: The first page is coherent. The second page is not.
Subject-19-H: I know.
Gunn: What happened?
Subject-19-H: I was trying to write the part that mattered.
Gunn: In English?
Subject-19-H: Yes. It was perfectly clear until I put it down.
Gunn: Can you give it verbally?
Subject-19-H: [pause] Not in English. I think I can say it another way, but I don't know how I know the other way.
Gunn: Try.
Subject-19-H: [in Gaelic] It told me I had crossed correctly. I think I want to know what else it would call correct.
Researcher note: this is the fourth subject account to describe the address as a primary motivation for a re-entry request. Subject-19-H has not developed persistent PLNS-L, but the second page of the submitted request meets criteria for transient English-output collapse. The three previously approved review-pending requests should be re-evaluated in light of this pattern. Recommendation forthcoming.
Addendum 7891-G — Site-71 Medical Summary: Post-Liminal Neurological Syndrome
Prepared by Site-71 Medical, with commentary from Dr. A. Gunn and the Foundation Linguistics Division.
Post-Liminal Neurological Syndrome (PLNS) is a provisional umbrella designation for delayed neurological presentations documented following exposure to SCP-7891. PLNS is not currently considered a single disease process. It is a filing category for symptoms that appear to share an association with threshold exposure and do not fit ordinary post-anomalous stress, trauma, infection, or neurodegeneration.
The three recognised PLNS presentations are as follows:
PLNS-T: transient temporal amnesia. Subjects lose access to discrete intervals surrounding entry, dawn-resolution, or return from displacement. The missing interval may range from several minutes to several hours. Subjects do not present as dissociated, psychotic, or confused outside the affected interval. They often know that time has passed and show appropriate concern when told what occurred during the missing period. The amnesia is temporal rather than personal. Subjects know who they are. They do not know where the missing time went.
PLNS-L: selective English-output failure with Gaelic substitution. Affected subjects retain English cognition, English comprehension, and in most cases the ability to read English silently. Internal monologue is reported as English. However, all externalised English output — spoken, written, typed, or signed — becomes severely disordered. Speech presents as phonological collapse with unstable substitutions, false starts, and non-lexical sound clusters. Written and typed English present as catastrophic dyslexia and dysgraphia: words deform, spelling fails to remain stable between attempts, and syntax breaks down into unrecoverable strings. Subjects are frequently aware that the output is wrong but cannot repair it.
PLNS-L subjects communicate coherently only in Scottish Gaelic or in Gaelic-adjacent historical forms. This fluency is not dependent on prior knowledge. Of the confirmed PLNS-L cases, only one subject had conversational Gaelic before exposure. Two had minimal passive familiarity. Three had none. All six produced grammatical Gaelic during presentation. When asked what language they are thinking in, all six answered, in Gaelic, that they were thinking in English.
English prompts may be understood by PLNS-L subjects, but English should not be required for routine communication. Repeated attempts to force English-language output have produced distress and no useful clinical data. This is the origin of the current containment requirement that Gaelic be used for external communication with PLNS-L subjects.
PLNS-CJD: progressive anomalous prion disease. PLNS-CJD clinically resembles Creutzfeldt-Jakob disease, including rapid cognitive decline, myoclonus, cortical ribboning, and abnormal prion-seeding assays. It is molecularly distinct from all known human prion strains. In both confirmed cases, first symptoms appeared after third-arch exposure, with onset at 84 and 117 days respectively. Neither subject developed PLNS-L. One presented with PLNS-T at return.
Neuropathological samples contain a misfolded prion protein conformation provisionally designated PrP-7891. PrP-7891 has not been observed in control personnel, local fauna, soil, lichen, standing water, or stone scrapings from the archway. Transmission studies are suspended pending Ethics Committee review. No secondary case has occurred among medical staff, researchers, or close contacts.
The relationship between the three PLNS presentations is unknown. Site-71 Medical currently treats them as separate expressions of post-liminal exposure. Dr. Gunn has objected to the word separate in this context.
Case Note: Subject-21-T
Subject-21-T crossed the first arch during the active period and experienced a moderate threshold state. Morphological changes resolved at dawn. Nine days later, the subject contacted Site-71 Medical to report that written English had become "slippery." Initial remote assessment dismissed this as stress. During in-person review, the subject read an English paragraph silently and answered comprehension questions correctly by pointing to multiple-choice options. When instructed to read the paragraph aloud, subject produced no intelligible English. When provided with a pen, subject attempted to write the sentence I can understand this but I cannot say it. The output could not be transcribed.
Dr. Ealasaid asked, in Gaelic, whether the subject understood the question. Subject-21-T answered: Tha mi ga thuigsinn sa Bheurla. Chan urrainn dhomh a thoirt a-mach sa Bheurla. Translation: "I understand it in English. I cannot bring it out in English."
Subject-21-T had no documented Gaelic proficiency before exposure.
Case Note: Subject-14-C
Subject-14-C crossed the second arch at the waning phase and did not initially present abnormal findings. Thirty-one days later, the subject was found unable to complete a written English workplace report. The report began coherently and degraded over two paragraphs into repeated malformed strings. Subject-14-C became increasingly agitated, insisting that the intended text was "still normal inside."
During evaluation, Subject-14-C communicated in fluent Gaelic for 19 hours before English output partially returned. The subject retained no conscious memory of learning Gaelic and described the temporary loss of English output as "standing in a doorway and watching the words refuse to cross."
Case Note: Subject-03-P
Subject-03-P crossed the third arch in 2020. Return occurred at dawn with no visible injury. Subject reported a 43-minute temporal gap and described the missing interval as "already spent." At 84 days post-exposure, subject developed tremor, sleep disturbance, and rapidly progressive cognitive impairment. MRI findings suggested prion disease. CSF assay was positive for abnormal seeding activity. Sequencing did not match known Creutzfeldt-Jakob disease variants.
Subject-03-P died 148 days post-exposure. Autopsy was conducted under prion-containment conditions. PrP-7891 was identified post-mortem. No secondary transmission has occurred.
Subject-03-P's final intelligible statement was in English and is preserved in the medical archive. It did not reference SCP-7891, the address, the moon, the wolf, or the clearing. This is included because not every case becomes meaningful in the way researchers want it to. Sometimes the harm is simply harm.
Linguistics Division Note
The Gaelic produced by PLNS-L subjects is not uniform. Some subjects produce contemporary Scottish Gaelic. Others produce archaic forms or dialectal features consistent with no single living community. Two recorded utterances contain constructions closer to Old Irish than modern Gaelic, though both were mutually intelligible to consulting linguists after review.
Affected subjects do not translate from English to Gaelic in the ordinary sense. Response times are consistent with native production, not second-language search. Subjects often express surprise or distress after hearing themselves speak. Several have asked whether they are "still saying what they meant." In all tested cases, semantic content is preserved. The thought remains English. The exit is Gaelic.
Dr. Gunn's Attached Comment
I dislike the medical neatness of this appendix. It makes the syndrome look like three bullet points. It is not three bullet points when someone is crying because they can read the sentence in front of them and cannot say it aloud without destroying it.
Still, the appendix is necessary. People have come back from SCP-7891 with missing time, with an impossible second language, and in two cases with a fatal prion disease. That has to sit in the same file as the relief, the address, the mirror, and the old records that insist the place is not cruel.
The uncomfortable part is that both sets of facts appear true.
— Dr. A. Gunn
Document last reviewed: [DATE REDACTED]
Classification: Euclid
Site-71 archive reference: 7891-PRIMARY