Extended Matrix

Extended Matrix Development Projects

← Back to Roadmap
DP-29

Narrative Units (EMN)

New Domains v2.0 Thesis candidate yEd paletteEMtoolss3Dgraphys3D config (rules)Heriverse

Description

Extended Matrix has been documenting what was built for over a decade — the strata of past construction, the network of physical and virtual stratigraphic units, the chain of reconstructive claims and the documentary sources behind each one. What it has not, until DP-29 picks up Paola Frigerio’s master-thesis prototype, formalised is what was lived inside what was built: the dialogues spoken in a Roman forum, the ritual acts performed at a temple threshold, the ambient sounds of a market portico, the social register of an exchange between two merchants. These are the patrimonio culturale immateriale — intangible cultural heritage as defined by the 2003 UNESCO Convention (ratified in Italy in 2007): oral traditions and expressions, performing arts, social practices and ritual events, knowledge about nature and the universe, traditional craftsmanship. The Frigerio thesis identified this as the conceptual gap the EM language did not yet address, and proposed the EMN — Extended Matrix Narratives — as the extension that closes it.

The thesis (Università di Padova, Corso di Laurea Magistrale in Scienze Archeologiche, A.A. 2022–2023, relatore E. Demetrescu, correlatrice C. Previato, developed in part during a placement at CNR ISPC Roma with Eva Pietroni’s mentoring) settled on a single methodological framing that drives every design decision below: the EM, extended by EMN, becomes the diario di lavoro held by the academic component of a development team, controlling coherence and validating the contents the creative component iterates on. The instrument stops being purely a documentation tool for the reconstruction researcher and becomes the shared notebook between two communities — academia and creative industries — collaborating on edutainment projects that go beyond either pure dissemination or pure entertainment. Frigerio’s Capitolo I interviews surfaced this gap from both sides: academics worry about non snaturare le informazioni, creatives have commercial and production constraints, and neither has a methodology that lets them author in the same artefact while keeping their own contracts.

The EMN closes the gap by adding one concept to the EM language: the Unità Narrativa (UN). UN behave like USV — they are graph nodes with validated properties, they reuse the existing EM symbology (no parallel notation), and they connect to the rest of the graph via the standard connectors. Two granularities share the class. General UN hold the framing layer — a narrator’s description, ambient sound, environmental cue — while specific UN hold the performed layer — a single dialogue line, a gesture, a ritual act. Either kind carries the four-axis property schema the thesis singled out for dialogue: tono (tone, situational), registro (linguistic register, documentary), contenuto (content, documentary), tempo (delivery rhythm, situational). The split between documentary and situational is the working-journal contract at the level of individual properties: registro and contenuto require at least one attested source, tono and tempo may be creative choices, and the EM records which axis carries which kind of justification so neither side has to take the other’s word for it.

UN attach to the rest of the graph through three classifying cases the thesis identifies in Capitolo IV.3. Caso singolo: one US/USV (a character, an object) connects to one UN (a monologue, a narration, a single line). Sistema o scena: several US and USV jointly host a scene UN that wraps multiple actors, interactions and dialogues — the EMN’s nearest neighbour to a stage direction. Racconto dentro al racconto: a UN opens a further narrative layer when a scene cannot be flattened — a story-in-story representation that lets a meta-narrative deepen a scene without collapsing the outer one. This last case is where the EMN earns its name as narratives (plural): the graph is structured to hold multiple narrative depths, not a single flat playback.

Characters in the EMN are USV. Frigerio’s thesis is explicit on this: ‘Questi ultimi vengono, in questo modo, ad essere rappresentati come USV’. Characters carry the same property bundle as any virtual stratigraphic unit — clothing, appearance, posture, behaviour — validated by documentary and creative sources in the standard EM paradata chain. The thesis deliberately defers the detailed character class work to the collegi dell’Università di Alicante, the team that has been developing the apparenza / movenze / comportamenti dimension under Sofia Pescarin’s group. DP-29 follows the thesis exactly here: it does not unilaterally introduce a new em:Character class. The character class refinement is collaborative future work, threaded with Alicante before any class addition lands in the EM language. For 2.0, UN reference characters as USV via the standard connectors and inherit whatever character formalisation the Alicante collaboration produces.

The thesis validated the EMN on two case studies, both worth preserving as canonical examples when DP-29 documentation lands. ‘Due mercanti a Nora’: a colloquial Latin dialogue between Claudius (a sea-trading ship’s captain) and Faustus (a local merchant interested in Campanian wine) in the Roman forum of Nora (Sardinia), late republican / early imperial. The Plautine corpus — chiefly Asinaria, Aulularia and Pseudolus — supplies the registro validation: the dialogue uses Plautine colloquialisms, attested expressions, the documented register of commercial exchange. ‘Archie il robot e i tre piccoli monelli’: the Segni acropolis CNR ISPC web-app (church of San Pietro / temple of Iuno Moneta), where the user is guided by two robot archaeologists (Archie and Archibald, beginner and expert variants) and the past-layer scene shows a priest reproaching three children playing noisily between the temple columns. Source-wise this is a different problem — the Plautine corpus is silent on temple scenes — so the thesis demonstrates how the EMN works across documentary sources by case. The Nora dialogue is intrattenente-first with documentary anchoring; the Segni scene is espositivo-first with the same EM machinery underneath.

The Heriverse playback layer (DP-11) is where the EMN repays the work in the visitor’s experience. With UN as first-class entities, a published EM model can drive an in-browser runtime that surfaces the intangible-heritage layer on top of the reconstructed 3D scene. MVP: subtitle overlay synchronised with UN ordering and tempo, each line one click away from its documentary attestations so the diario di lavoro becomes visible to the public. Character animation, TTS narration of registro-validated lines, and ambient sound design of general UN are 2.x enhancements, coordinated with the Alicante character work and with separate DPs around music (DP-23) and dance (DP-24) reconstruction. The point is not to turn the EM into a game engine — it is to make the intangible heritage layer auditable and citable in the same way the material layer already is.

Why this is in the EM 2.0 horizon rather than 1.7. The EMN is a substantial language extension, and Frigerio’s thesis was already theoretical-complete in 2022–2023; what it needs is the surrounding EM language infrastructure to mature, which the 1.6 / 1.7 cycle has been delivering. Canonical edges (DP-62) so UN ↔ USV roundtrip cleanly through GraphML + pyArchInit. Multilingual labels (DP-63) so registro controlled-vocabulary entries can be authored in Italian / English / German / Greek. Property inheritance + Combiner premise (DP-66) so the four UN axes plug into the standard paradata chain. Propagative metadata (DP-32) so author / license / embargo on UN behave identically to every other stratigraphic node. Three language-cycle iterations later, the EM is finally in shape to host UN cleanly. Once 1.7 closes, DP-29 ships as a 2.0 cornerstone — Extended Matrix moves from documenting the past to performing it, with the academic and creative communities iterating in the same artefact.

Open decisions for the review call (seven points). (1) UN as USV-shaped node with unita_tipo=UN vs new top-level class — proposal: USV-shaped, distinguished by unita_tipo, so canonical-edges roundtrip stays trivial and the EM legend stays unchanged. (2) Character class refinement — not for DP-29 alone to settle; threaded with Alicante as a separate DP when the Alicante side is ready. (3) Three-case taxonomy as narrative_case attribute (singolo | sistema_scena | racconto_dentro_racconto) driving the EM Tools panel layout — proposal: yes. (4) Four UN properties (tono, registro, contenuto, tempo) modelled as discrete fields, each with its own paradata chain via DP-66, validation rule requiring documentary source on registro + contenuto — proposal: yes. (5) Heriverse playback: subtitle-overlay MVP, animation + TTS deferred to 2.x coordinated with Alicante. (6) registro as DP-09 controlled vocabulary with DP-63 multilingual labels — proposal: yes. (7) AI / NLP assistance (Frigerio Capitolo V.1) — adjacent thread, separate DP after 2.0 ships.

Status

Prototype (thesis) — needs language formalisation

Target EM Version

2.0

Impacts

yEd paletteEMtoolss3Dgraphys3D config (rules)Heriverse

Components

  • **UN (Unità Narrative)** as USV-shaped graph entities. Frigerio's thesis defines UN as nodes that behave *like* USV: they carry properties, and those properties are validated through documentation just as USV properties are. Two granularities live in the same class: *general UN* (an external narrator's description, ambient sounds, environmental cues — the framing layer) and *specific UN* (a single line of dialogue spoken by one character, a gesture, a ritual act — the performed layer).
  • **UN properties for a dialogue line** — four named axes from the thesis (IV.3): `tono` (tone, situational), `registro` (linguistic register, documentary-validated), `contenuto` (content, documentary-validated), `tempo` (rhythm / speed of delivery, situational). Two of the four (`registro`, `contenuto`) are anchored to documentary attestation; two (`tono`, `tempo`) are creative choices made by the writing team. The split is the working-journal contract: the academic side validates the documentary axes, the creative side owns the situational axes, and the EM graph records which axis carries which kind of justification.
  • **Three case taxonomy** (Frigerio IV.3) for how UN attach to the rest of the graph: *(i) caso singolo* — one US/USV (a character, an object) connects to one UN (a monologue, a narration, a single line); *(ii) sistema o scena* — several US and USV jointly host a scene UN that wraps multiple actors, interactions, dialogues; *(iii) racconto dentro al racconto* — a UN opens a further narrative layer (a new sub-plot) when a scene cannot be flattened, allowing a story-in-story representation without collapsing the outer narrative.
  • **Characters as USV with deferred class refinement.** The thesis is explicit: ‘Questi ultimi vengono, in questo modo, ad essere rappresentati come USV’ — characters are USV, carrying properties like clothing, appearance, posture, behaviour, validated by both documentary and creative sources. Detailed character-class work is reserved by Frigerio to ‘i collegi dell'Università di Alicante’ (the Alicante group working on apparenza / movenze / comportamenti). DP-29 follows the thesis: it does not unilaterally define a new `em:Character` class. The character class question is open collaborative work to be threaded with Alicante before any class addition lands in the EM language.
  • **EM as diario di lavoro** — the methodological framing the thesis settles on: the EM (extended by EMN) is the academic side's working journal, held to control the coherence and validate the contents that the creative side iterates on. The EM stops being purely a documentation instrument for the reconstruction researcher and becomes the shared notebook between two communities — academia and creative industries — toward edutainment projects that go beyond either pure dissemination or pure entertainment.
  • **Intangible cultural heritage anchor (UNESCO 2003 / IT ratification 2007)** — DP-29's deepest conceptual root: the patrimonio culturale immateriale convention defines oral traditions, performing arts, social customs, ritual events, beliefs, traditional craftsmanship as cultural heritage on equal footing with the material side. The EMN's purpose is to bring this layer into the EM workflow so that an EM-documented reconstruction can record not only *what the building looked like* but *what was said in it, by whom, with what register, under what ritual, with what surrounding ambient sounds and customs*.
  • **Reuse of the existing EM symbology, not a new notation.** Frigerio explicitly uses ‘le simbologie base dello stesso strumento e presentate in Figura 8’ — the EMN reuses the EM node legend, no separate visual language. UN render as USV-shaped nodes with documented type tags; edges to US/USV use existing EM connectors. This is a deliberate choice: the EMN is an extension *within* the EM language, not a parallel one.
  • **EM Tools authoring affordance** — a Narrative Units panel parallel to the Activity / Epoch / TimeBranch managers (DP-12 + DP-43). UN appear by case (singolo / sistema-scena / racconto-dentro), each row showing actor (USV character), the four-axis properties, and the document attestations behind `registro` + `contenuto`. The panel lets the academic side query *which UN have unvalidated documentary axes* — that is the validation worklist the diario di lavoro framing implies.
  • **Two case studies in the thesis** (Capitolo IV) to honour in the documentation when DP-29 lands: (1) *‘Due mercanti a Nora’* — a dialogue between two merchants in the Roman forum of Nora (Sardinia), late republican / early imperial, with the Plautine corpus (Asinaria, Aulularia, Pseudolus) supplying the colloquial register validation; (2) *‘Archie il robot e i tre piccoli monelli’* — the Segni acropolis CNR ISPC web-app (San Pietro church / Iuno Moneta temple) using a robot-archaeologist narrator (Archie and Archibald) plus a temple scene with a priest and three children, more expository than the Nora case but still anchored on documented Roman religious culture.
  • **Heriverse playback** (DP-11 integration) — once UN are first-class EM entities, the Heriverse viewer can render them as a runtime layer over the published 3D scene. MVP affordance: subtitle overlay synchronised with UN ordering / duration; later, character animation (collaboratively with the Alicante team), TTS narration of dialogue lines, sound design of `general` UN (ambient cues). Each UN's documentary attestations stay one click away from the visitor — the diario di lavoro becomes a visible feature of the public-facing experience.
  • **Future-work hooks from Frigerio's Capitolo V.1**: AI / NLP assistance for compiling the documentary dossier and for first-draft dialogue generation (the thesis appendix already records a ChatGPT experiment); ethical guardrails around training-data licensing and creative-labour displacement; VR / AR delivery formats as virtual-museum experiences. These are signposted in DP-29 as adjacent threads, not in-scope for the 2.0 language formalisation.
  • **RDF / SPARQL** surface enabled by UN as first-class entities: *‘list all UN attached to a given USV character ordered by the case taxonomy’*; *‘list UN whose `registro` is documented vs creative-only’*; *‘trace which documentary sources validate which UN content axes’*; *‘count UN per Activity / Epoch / TimeBranch to find the narrative density of a reconstruction’*. Maps to the v1.6 RDF Export layer once UN is declared in `em.ttl`.

Key Study

Frigerio (2022–2023). EMN as theoretical extension validated on two case studies — Nora forum (Plautine register source) and Segni acropolis (CNR ISPC web-app). Thesis status: theoretical with prototype application; needs further testing on full projects and language-level formalisation in EM, in coordination with the Alicante team for the character class.

Notes

Original prototype lives in Paola Frigerio's master thesis *Extended Matrix Narratives: informare e intrattenere col passato*, Università degli Studi di Padova, Corso di Laurea Magistrale in Scienze Archeologiche, A.A. 2022–2023, relatore E. Demetrescu, correlatrice C. Previato. The thesis was developed in part during a placement at CNR ISPC Roma, with Eva Pietroni's mentoring (acknowledged in the thesis preface). DP-29 carries the prototype into the EM 2.0 language formalisation. Open decisions to thread at review: (1) UN as USV-shaped sibling vs sub-type of a new top-level class. Frigerio leans USV-shaped (same legend, same node behaviour); proposal: keep UN USV-shaped at language level, distinguish by `unita_tipo` value so canonical-edges round-trip stays trivial. (2) Character class refinement — *not for DP-29 alone to settle*. The thesis defers character work to the Alicante team (Pescarin et al.); DP-29 commits only to treating characters as USV with documented + creative property axes and to threading the class refinement with Alicante as a separate DP when the Alicante side is ready. (3) The three-case taxonomy lives where? Proposal: as a `narrative_case` attribute on the UN node (`singolo` | `sistema_scena` | `racconto_dentro_racconto`), driving the EM Tools panel layout but not changing the underlying node class. (4) The four UN properties (`tono`, `registro`, `contenuto`, `tempo`) — modelled as discrete fields each carrying its own paradata chain? Proposal: yes — each property uses the standard EM Property → Extractor / Combiner paradata structure (DP-66), with `registro` and `contenuto` requiring at least one documentary source, `tono` and `tempo` allowed to be creative-only. The validation rule mirrors Frigerio's working-journal contract. (5) Heriverse playback (DP-11): subtitle overlay MVP, TTS / character animation deferred to 2.x coordinated with Alicante. (6) The thesis prototype is Plautine Latin; DP-29 generalises to any historical register via the Vocabulary Project (DP-09) so the cross-project register catalogue can serve any language community; multilingual labels for register entries via DP-63. (7) AI / NLP assistance (Capitolo V.1) — adjacent thread, not in-scope for DP-29 language formalisation; revisit as a separate DP after 2.0 ships. Cross-refs: DP-23 (Music Reconstruction) and DP-24 (Dance Reconstruction) — sibling temporal-performance reconstruction DPs that share the intangible-culture extension framing; DP-28 (Virtual SU Pattern) — the cultural-pattern reasoning that UN dialogue registers can inherit from; DP-12 (Temporal Branches) — alternative narrative readings as exclusive / concatenated branches (the working-journal exposes ‘Plautine register vs Terentian register’ as two TimeBranches the academic + creative pair iterate on); DP-43 (Group Nodes — Activity in particular) — a scene-case UN can live inside an Activity that holds its actors and props; DP-11 (Heriverse) — the playback runtime; DP-30 (Reconstructive Elements — props and characters with microstratigraphy and hypergraph link) — adjacent and shares the character question with Alicante; DP-09 (Vocabulary Project) — home of the `registro` controlled vocabulary; DP-66 (property inheritance / Combiner premise) — UN property paradata structure reuses the standard chain; DP-32 (Propagative Metadata Resolver) — author / license / embargo on UN as for every other stratigraphic node. EM version note at the time of the thesis: EM 1.3 (current published); current EM at the time of DP-29 documentation (June 2026) is 1.6 stable + 1.7 dev — three language-cycle iterations later, the surrounding infrastructure (canonical edges DP-62, multilingual labels DP-63, property inheritance DP-66, propagative metadata DP-32) needed to host UN cleanly is finally in place.