Sub-People Inside a Single Person

The big idea, in one breath

Scott’s paper proposes that inside one named “part” of a person with Dissociative Identity Disorder, there can be several quieter versions of that same part, each carrying a slightly different bundle of memories, and each switched on by how safe the body feels at that moment.

The science of DID already explains, quite well, how a person switches between named alters: how Penny can give way to Penelope, who feels and behaves differently. What Scott noticed, over eight years of living alongside his partner, is something one level deeper and stranger.

Sometimes a presentation arrives that looks exactly like Penny, sounds exactly like Penny, and is completely certain she is Penny, and yet she does not remember something that Penny plainly knew. Everything about who she is stays intact. Only a specific slice of memory is missing, and it is missing so cleanly that she does not feel any gap at all.

Scott’s explanation is that a named alter need not be a single fixed person inside. It can behave more like a broad template for a person, a kind of outline of “who Penny is”, and that outline can be filled in more than once, producing several versions of Penny. They share the same name and the same heart. What they do not always share is the same set of episodic memories. He calls these inner versions sub-alters, and the whole idea intra-alter substructure, which simply means “structure inside an alter”.

The reason this matters is small but real: clinicians and partners have run into this experience for decades, and the formal research literature had never given it a name or an explanation. Scott’s paper does both. That is its contribution. It does not claim to have proved the idea with experiments; it builds the idea carefully out of work that is already accepted, and then shows what it would mean for treatment if it were true.

The summary at the front of the paper

Every scientific paper opens with an abstract: a single dense paragraph that acts like a trailer for the whole film. Scott’s abstract states, in compressed form, the five things the paper will do.

It says, first, that current theories explain switching between alters but do not explain a particular pattern: a presentation that wears an alter’s name and whole sense of self while reaching a different store of memories, with no outward announcement that anything is different. Second, it names the new idea: intra-alter substructure. Third, it lists the four established bodies of research it will weave together. Fourth, it previews the three-layer picture and the “safety decides who comes forward” mechanism. Fifth, it draws out what this would change about treatment, ending with a fresh way of thinking about what healing even means.

The abstract also mentions one tool by name, the Beach Safety Hierarchy Assessment Scale. That is Scott’s own earlier instrument, which is why his name appears as the source. We will come to what it measures further down.

The opening: the thing nobody had named

The introduction begins with a scene that anyone close to a person with DID will recognise. You are talking with a known alter. The name is the same, the self-concept is the same, the warmth toward you is the same. And then a single remark gives it away: she refers to something the two of you have shared dozens of times as though she has never encountered it before. Who she is remains completely intact. What is missing is one bounded slice of remembered experience.

Scott states plainly that this experience “has no direct home in the peer-reviewed DID literature”. In other words, the field had the experience but not the vocabulary. He then names the existing theories that come close without quite arriving, and we will meet each of them properly in Part 5. His central claim is that a named alter can act as a broad template whose versions are switched on, below the level of anyone’s conscious choice, by the body’s moment-to-moment reading of safety.

Then he offers the reframe that colours everything after it, and it rests on the work of one researcher in particular.

The paper closes its introduction by laying out its four steps: review the existing theories and find the gap, describe the new model in full, present the real-life observations, and draw out the consequences for treatment. The honesty of that structure matters. Scott tells you up front that the personal observations are offered as illustration, not as proof.

Is this condition real? The honest argument

Before proposing anything new, Scott does something careful: he acknowledges that the reality of DID itself is debated, and he sets out where he stands. There are two long-standing camps.

The first, the trauma model, holds that DID is a genuine condition caused by severe early trauma, exactly along the developmental lines Putnam described. The second, the sociocognitive model, holds that DID presentations can be shaped by suggestion, by a person’s tendency toward rich fantasy, by films and books, and even, unintentionally, by therapists themselves.

Scott sides with the trauma model, and he does not just assert it. He brings two specific pieces of evidence that are widely regarded as among the strongest in the field.

He also references Brand and colleagues, whose reviews of treatment-outcome studies sit firmly within the trauma-model tradition and show that trauma-focused treatment for dissociative disorders produces real improvement. Together, Dalenberg, Reinders, and Brand let Scott establish his footing honestly: he has named the opposing camp, and he has shown why he reasonably stands where he stands.

The four bodies of work the idea is built on

Scott is careful to present his idea as an extension of accepted work rather than a break from it. He leans on four established bodies of research, plus the clinical writing of one pioneer. This is the longest stretch of the paper, so we will take the building blocks one at a time.

Building block one: structural dissociation

Building block two: Putnam’s discrete behavioural states

We have already met Putnam’s developmental idea in Part 3, and it returns here as a full building block. The key addition is that Putnam’s model is naturally layered: the same weaving process that would have joined a child’s simple states into one self would also have organised smaller substates into larger ones. DID, in this view, is the failure of that organising process at several levels at once. That layered quality is precisely what lets Scott propose a further level of structure inside an alter without contradicting Putnam. The 2022 chapter by Loewenstein and Putnam restates this developmental account for a current clinical audience. Putnam also passed along a vivid observation from the next source we meet, Richard Kluft: that “all alters are not the same all the time”.

Building block three: Kluft’s clinical observations

Building block four: schema mode theory

This is the bridge that carries Scott from the established literature to his own idea, so it is worth slowing down for.

Building block five: a part he deliberately set aside

The trickiest objection: memory

Scott’s whole idea rests on memory differing between versions of one alter. So he stops and deals head-on with the most contested question in the entire field. This is the most intellectually demanding part of the paper, and also the most impressive, because he turns a potential weakness into a careful strength.

Here is the clever move. Faced with strong evidence that memory is not truly sealed between alters, Scott does not insist that it is. He changes the shape of his claim to fit the evidence. He proposes that the memory differences between his sub-alters are not a case of one version being physically unable to access a locked vault. They are a case of state-dependent retrieval: the memory exists somewhere in the wider system, but it is hard to reach from the particular bodily and emotional state the current version is in.

The new idea itself: an alter as an umbrella

Now Scott states his proposal in full. He keeps everything the existing theories say about a named alter being a real, distinct identity, and adds one thing on top: a named alter is also a broad template. It holds the stable core of who this identity is (the roles, the values, the loves, the way of relating) without fixing in advance the precise memories or exact mood active at any given moment. The stable core is shared by every version. The lived memories are not.

From this follows the most thought-provoking claim in the paper, and it is worth reading slowly. A sub-alter does not know it is a sub-alter. The reason is quietly logical. We feel ourselves to be one continuous person on the basis of our sense of identity, our values and attachments and name, and not on the basis of having a complete, gap-free reel of every episode of our lives. A sub-alter carries the full sense of identity of the named alter. It does not carry the episodes during which other versions were present. And here is the key: those missing episodes do not feel missing. There is no sense of a gap, no feeling of lost time, because this version was simply never there for them. From the inside, the sense of self is complete and seamless. She experiences herself as Penny because, at the level of identity she can actually feel, she truly is Penny.

From the outside, then, only one kind of evidence ever gives it away: a memory gap. A reference to something long-established met as though brand new. That is the single thread that, when pulled, reveals the whole hidden structure.

The three layers, and the iceberg

Scott proposes that these sub-alters are not all alike. They sit along a spectrum, which he divides, for description, into three layers. The layers differ in three ways at once: how easy the version is to detect from outside, how aware it is of its own difference, and how complete its memory gap is. He stresses that these are points on a slope rather than three separate boxes.

Layer One: the named, known parts. These are the recognised alters of the system, the ones with names, the ones that take part in the system’s inner life. They are known to the person, often to a partner, and to a clinician. They know they are part of a larger whole. Penny, Penelope, and Penny Love all live here.

Layer Two: detectable sub-alters. These are versions stable enough to carry a small, repeatable signature, a particular way of holding the mouth, a posture, a vocal quality, that reliably differs from the parent alter’s baseline. They are not named within the system and do not sit at its inner table. Only an intimate observer with years of accumulated baseline can spot them, and even then barely. They may carry a partial awareness of their own difference, and their memory gap is partial rather than total.

Layer Three: covert sub-alters. These are the deep ones. They are indistinguishable from the parent alter to outside observers, to the rest of the system, and to themselves. They show no outward signal of difference at all. Their memory gap is the most complete: something that was well-established and emotionally important to the parent alter is simply absent, felt not as forgotten but as having never happened. They can only ever be detected afterwards, by the gap they leave behind. The earlobe story in Part 11 is the clearest example.

What decides who comes forward

If there are several versions of one alter, what chooses which version is present at a given moment? Scott’s answer is the heart of the model’s machinery, and it draws on the work of Stephen Porges.

Scott then draws out a consequence that gives the whole paper its tenderness. The person who provides steady safety, an intimate partner above all, is doing something physical, not just emotional. Their own calm, carried in a warm voice, a settled face, and reliable presence, is read by the other person’s nervous system as a safety signal. He calls this being a co-regulator. It means a partner is, without trying and without either person choosing it, helping to determine which versions of their loved one are able to come forward. A dysregulated, frightening environment summons the defended versions. A calm, predictable, warm one summons the open ones.

Safety as the key that unlocks the deeper layers

Two further ideas complete the model.

The first is permeability. Sub-alters are not sealed jars. When two named alters live close together in daily life, sharing the same roles of partner and parent, a version of one can slowly take on some of the qualities and skills of the other, through being repeatedly present in the same overlapping situations. The result is a state that feels like both at once. This is where Edwards’ idea of blended modes, met earlier, does its work, and it leads to a gentle reframing of what healing might mean: not the forced merging of separate identities, but the gradual softening of the boundaries between versions as they come to share more and more of the same life. Under sustained safety, the parts grow more alike simply by living the same days together.

The second is the map of safety itself, which is Scott’s own contribution.

The five levels map onto the model like this:

The thread tying this together is the model’s most practical claim: you do not summon the deeper versions deliberately. You make the room safe enough, for long enough, that they come forward on their own.

The real-life observations

This is the part of the paper closest to home. Scott illustrates the model with five observations gathered over more than eight years of living alongside his partner, who has a confirmed diagnosis of DID. He is openly honest, in the paper, about what this evidence is and is not, and that honesty is itself worth seeing.

The earlobe (a Layer Three covert version). Over the years, his partner had often remarked, affectionately, on a small wrinkle on his earlobe. It had become a little marker of intimacy between them. One evening, with no change at all in voice, posture, or warmth, she looked at him and said she had never noticed that wrinkle before, as though seeing it for the first time. The next morning she had little memory of the evening. Everything that made her herself was intact. One specific, well-worn shared memory was simply absent. This is the covert version revealed only by the gap it leaves.

“I see you” (a Layer Two detectable version). Over years, Scott had noticed a subtle, repeatable detail, a particular set of the mouth, that reliably differed from his partner’s usual presentation while everything else stayed the same. One day, noticing it again, he hugged her and made his quiet observational check. She laughed and said, “You don’t know me, ha.” He replied, “I see you.” Three things stood out: she showed partial awareness that he had noticed something; her response was playful rather than guarded, a sign of safety rather than threat; and the marker was so subtle it had taken years of baseline to spot.

“I feel a little of both” (blending across versions). Just after that exchange, a shift occurred toward Penelope, the higher-energy, outgoing alter, yet she carried on with the gentle domestic caregiving more typical of Penny. Asked which one she felt she was, she said she felt a little of both. This is the permeability and blending from Part 10, observed directly: two normally distinct identities, sharing the same daily roles, softening into one another.

The names (memory varying while identity holds). This is the most clear-cut observation. Different parts address Scott differently. Penny calls him Scott. Penny Love, with a softer voice, also calls him Scott. But LittlePK, a child part anchored to around 1986 at roughly age six, calls him by his full name, Scott David Beach, the way a young child would formally address an unfamiliar grown-up. Every part knows, at the level of identity, that he is her safe person. What differs is the lived history: the child part has not accumulated the years of easy closeness that make a first name feel natural. The same sense of self, a different store of episodes, shown in something as simple as a name.

The iceberg, from inside. During a conversation with Penny Love about how many parts attend the system’s nightly inner meetings, she reported hearing an internal voice say that those are “just the tip of the iceberg”. Remarkably, this image of a small visible surface above a far larger hidden structure arose from within the system itself, matching the three-layer picture Scott had arrived at on theoretical grounds.

What this could mean for treatment

If the model is right, it changes where a therapist aims. Scott draws out several consequences, all of them gentle and none of them requiring a brand-new technique.

The reach problem. If the named alters are only the visible surface, then treatment that works solely with named alters may be reaching only Layer One. The negotiated agreements, the processed memories, the relational repair, all of it may be landing in the surface layer, while the deeper versions keep coming and going, carrying their own memories, untouched. Scott is careful to say this is not a criticism of existing treatment and does not mean its results are unreal. It means those results may be incomplete in a way the field has not had the words for.

A different target. The answer he proposes is not a new method but a new target. Instead of aiming at the named alter, aim at the relational and bodily environment in which that alter appears. Make the environment safe enough at the right level, and which versions are available to come forward changes on its own. In his words, you do not call the deeper parts into the room; you make the room safe enough that they come forward themselves.

The acknowledgment principle. The “I see you” moment suggests that simply, warmly noticing a version that has never been acknowledged may carry real weight. But Scott is firm about the line between two ways of doing this. Gently observing “I notice something different today” is appropriate. Announcing “you may have sub-alters within your alters” is not, because naming structures for someone can accidentally create the very thing it claims to find. This is the oldest and most serious caution in the field, and he respects it.

A softer idea of integration. Finally, the model reframes the long debate about whether the goal of treatment is to merge all the parts into one or to help them live well together. If the boundaries between versions soften naturally under sustained safety, then integration need not be a forced merger at all. It can be the slow growing-together of versions that come to share more and more of the same life. Safety, in this view, is not the warm-up before the real work. Safety is the work.

What the paper honestly does not prove

Scott devotes a full, frank section to the limits of his own work, which is one of the most trustworthy things a researcher can do.

He acknowledges the one-person, one-observer problem in full: his observations may reflect the quirks of a single system, and he, as both partner and theorist, is not a neutral recorder. He acknowledges that no formal measurements were taken during those years, and lists the established instruments a proper future study should use.

He also concedes that the paper is theoretical: it does not establish how common this structure is, how the selection works at the level of brain cells, or that safety truly causes deeper access. He flags the risk of misuse if a clinician applied the idea heavy-handedly, and he leaves open whether the model extends to related, milder dissociative conditions.

Then he points forward. He proposes a careful interview guide (included at the back of the paper) for asking people with DID about these experiences without putting ideas in their heads; a method for gathering partners’ long-term observations; and, as the eventual gold standard, brain-imaging studies, modelled on the Reinders work met in Part 4, to test whether two versions of one alter produce genuinely different brain signatures. The simple naming-pattern observation, he notes, may be the easiest first thing to test across many people.

The quiet idea at the heart of it all

The paper ends by returning to the reframe it began with, because everything else rests on it.

DID is not a broken mind. It is a mind that completed exactly the process it was built to complete, under conditions that interrupted the final step.

The states did not fail to join because something was defective. They failed to join because the surroundings were not safe enough, for long enough, for the joining to finish. The named alters, the hidden versions in their layers, the inner meetings and the structure below the waterline, all of it is the shape a normal developmental process takes when it is stopped before completion.

And that points to a hopeful clinical conclusion. The task is not to fix what is broken. It is to provide now what was missing then: steady relational safety in which the interrupted process can quietly resume, even years or decades later. Given that, the system begins to move. Versions that come forward in safety carry more shared history than those that come forward in fear. The boundaries soften. A child part that once offered only the formal courtesy of a stranger learns, in time, to use a first name.

Scott’s closing line is that this is a quieter kind of integration than the literature usually describes, and perhaps one that is truer to what the developmental process was always trying to become.

Every source, in one plain list

For completeness, here is every work the paper cites, each in a single plain sentence, in alphabetical order by author, the way the reference list at the back of the paper is arranged. A few are marked as still being finalised, meaning the full publication details were to be confirmed before formal submission.

• Bachrach and colleagues (2023). A published account of one patient treated for DID with schema therapy over 220 sessions, with real improvements; a separate commentary on it was later retracted, but the case report itself stands.
• Bartlett (1932). A classic work showing that human memory is a reconstruction built around mental templates, one of the two scholarly roots of the word “schema”.
• Beach (2026). Scott’s own five-level scale of relational safety in a trauma-affected partnership, currently undergoing validation.
• Beker and colleagues (2024). A recent review of the evidence on memory between identities. (Full reference still being finalised.)
• Brand and colleagues. Reviews of treatment-outcome studies showing that trauma-focused treatment for dissociative disorders helps, supporting the trauma model.
• Braun (1986). Introduced the idea of a “fragment” (a part with a very narrow range) and the term “polyfragmented” for systems with very many parts.
• Carlson and Putnam (1993). Authors of the Dissociative Experiences Scale, the most widely used quick screen for dissociation.
• Dalenberg and colleagues (2012). A large, respected review that weighed the trauma and sociocognitive models against each other and found strong support for the trauma model.
• Dell (2006). Author of the Multidimensional Inventory of Dissociation, among the most thorough assessment tools for DID.
• Dell and O’Neil (2009). A major reference volume cataloguing how dissociated parts cluster into pairs, subgroups, and layers.
• Donath and colleagues (2025). A corrected re-analysis arguing that memory does transfer between identities even when amnesia is sincerely reported. (Full reference still being finalised.)
• Edwards (2022). Expanded the catalogue of schema modes and named the “blended mode” (two modes active at once) that Scott uses. (Publisher still being finalised.)
• Felitti and colleagues (1998). The landmark Adverse Childhood Experiences study, which measures the burden of early childhood adversity.
• Haddock (2001). A DID sourcebook that, with online community knowledge, describes an alter who has alters of their own, the lived-experience version of Scott’s idea.
• Howell (2005). Argues that the splitting seen in DID is an extreme form of a dissociative tendency present in all minds.
• Huntjens and colleagues (2002, 2006, 2012, 2014). A rigorous series of experiments showing that memory often transfers between identities even when a person sincerely reports amnesia.
• Huntjens, Rijkeboer, and Arntz (2019). Formally applied schema-mode theory to DID and built a treatment protocol; the direct idea Scott extends one level deeper.
• ISSTD Guidelines (2011). The field’s professional treatment guidelines, which use the term “subsystem”.
• Kluft (1988). The closest historical anchor; described layering, near-duplicate alters, and the mind “recopying itself selectively”.
• Linde and colleagues (2023). Compared schemas and modes across dissociative and personality disorders and healthy controls. (Volume and pages still being finalised.)
• Loewenstein and Putnam (2022). A recent restatement of Putnam’s developmental account of the dissociative disorders for clinicians.
• Merskey (1992). A leading sceptical voice arguing some DID cases are shaped by the diagnostic process itself.
• Nijenhuis and van der Hart (2019). A rebuttal arguing the schema-mode approach misreads structural dissociation theory. (Full reference still being finalised.)
• Piaget (1952). Used “schema” for the mental structures a child builds and adapts; the second scholarly root of the term.
• Porges (2004, 2011). Developed polyvagal theory and the idea of “neuroception”, the body’s automatic reading of safety, which supplies Scott’s selection switch.
• Putnam (1989, 1997, 2016). The developmental account of DID as an interrupted process; the foundation stone of the whole paper.
• Reinders and colleagues (2012). A brain-imaging study showing genuine DID states carry physical signatures that actors cannot fake; the paper’s strongest concrete evidence.
• Ross. Proposed loosening structural dissociation theory so a compartment can hold something as small as a single feeling. (Full reference still being finalised.)
• Rothschild (2000). Argued trauma is held in the body and that healing often needs a return to the original bodily state.
• Schwartz (1995). Founder of Internal Family Systems, a model Scott deliberately set aside because it is not a trauma-causation theory.
• Spanos (1994). A central figure in the sociocognitive view, arguing multiple identity can be a learned social role.
• Tulving and Thomson (1973). Established encoding specificity, the rule that memories return most easily in the state in which they formed.
• van der Hart, Nijenhuis, and Steele (2006). Authors of structural dissociation theory, the leading framework, which Scott extends to fill one gap it leaves.
• van der Kolk (1994). A widely known authority on how the body stores trauma, cited for state-dependent memory.
• Young, Klosko, and Weishaar (2003); Young and colleagues (2007). The founders of schema therapy and the Schema Mode Inventory, the source of the core schema-and-mode logic Scott builds on.

The sceptical view also draws on Lilienfeld and Lynn, named in the paper alongside Spanos and Merskey as voices in the sociocognitive tradition.