After Babel, The Machine?

Four readings of Gabriele Tergit — a comparative analysis

A description of how four readers — three AI, one human — constructed an understanding of the author Gabriele Tergit from different bodies of evidence. This is comparison, not ranking. The question of which reading is the truer or more faithful Tergit is left to human reviewers.

The four readers, and what each was reading

A structural fact that shapes everything below: the readers steeped very unevenly in Tergit’s own prose. A’s step-1 corpus notes run ~58,000 words; B’s ~48,000 (a line-by-line reading of five books, distilled into a twelve-point “voice signature”); D’s step-1 notes run only ~9,400 words, and D did not read Effingers itself until step 3. D leaned more of its weight on secondary framing and on the anglophone register exercise; A and B steeped far more deeply in the primary text. This asymmetry recurs below as an explanation for several differences.

1. Tergit-the-person

Where all four converge

On the biographical and temperamental core the four readings are strikingly consonant — and, for the two self-built reads (B, D), they converge with Sophie without ever having seen her.

The two Berlins, and “both sides of the door.” Every reader makes the move from the poor eastern Berlin of the father’s cable factory to the wealthy Tiergarten the foundation of Tergit’s eye.

This is the clearest case of independent recovery: B and D, from German sources, reconstruct exactly the social double-vision Sophie names — and tie it, as she does, to the court reporter’s refusal to treat a defendant as a specimen.

The name from the grille. All four give the Gitter/Tergit inversion and read it as self-definition. B and D both make it emblematic of the court reporter: B — “I spent the fat years watching the world through the courtroom railing”; D — “the iron grating, the courthouse railing where I would soon be standing every day.” (B and D both also know the arthropod-dorsal-plate pun, from German reference works; Sophie and A do not use it.)

The recording-band method, born of shyness. A and Sophie share the precise origin story — the timid reporter who would not take notes and so memorized the one decisive sentence. Sophie: “she felt self-conscious about taking notes and remembered the proceedings verbatim instead.” A: “I would not take notes in the room and learned instead to carry the one decisive sentence out in my head. That shyness became my method. I am a recording-band.” B independently arrives at the same instrument from the primary prose — “My brain is a recording-tape that holds the one decisive sentence” — and D at “I learned what people are by watching them under oath.”

The historian by training. B and D both foreground Meinecke and the Carl Vogt dissertation as formative (Sophie and A note it but weight it less). For B and D it is load-bearing: D — “I am a historian by training — Meinecke was my teacher… I want to see history happen”; B repeats the German “der Historie zusehen.” The German-built reads put slightly more structural weight here.

Exile as expulsion, not emigration. B and D independently seize on the Vertriebene-not- Emigrant distinction as identity-defining, in nearly identical terms — B: “I am not an Emigrant… I am one of the Vertriebenen, the driven-out”; D: “I do not say Emigrantin. I say Vertriebene. The distinction is exact.” Sophie records the expulsion vividly (the SA at the door, Heinz forbidding the maid) but does not make the Vertriebene terminology a centerpiece; A, following Sophie, does not foreground it either. This is a genuine emphasis-divergence: the German-built reads, drawing on Wagener, treat the word itself as a political-existential position; the Sophie-line reads treat the fact without the term. A small but real case where the German evidence base carries something the anglophone-source line does not.

The losses, stated flat. All four center Peter (killed in the Dolomites, 1964) and Heinz (1968) and read the flatness of Tergit’s report of them as itself her art. A: “I say these things flatly because that is how the worst things are true.” D gives a dated necrology of the whole circle (“the Untergrund of everything I write after 1945”). B: “Es ist alles sinnlos allein.”

Politics: the enemy of the generalization. The deepest convergence of all, holding across every evidence base. Sophie’s epitaph names “her profound skepticism of easy truths and overt ideology.” B makes it the literal center — “My greatest hatred is the generalization… The instant you say the Jew, the German, the woman, you have left the truth and entered the lie.” A: “I will not hate by category… Individual guilt, yes, named plainly… Collective liability, never.” D: distrust of “every idea that sets itself absolute, that asks me to become mass.” The even-handed hostility to both totalitarianisms is in all four; A and D both carry Tergit’s own line that beside the dictatorship of the proletariat the Nazis were “pure orphans” — a line that originates in Tergit’s own memoir, which is why both the Sophie-line and the German-line reads can hold it.

The Jewish question and anti-Zionism. All four give: rooted German Jew, Bildung as the form of Jewishness, blood-and-soil critique of Zionism learned from the German right, and post-Auschwitz ambivalence. Sophie: “She never was a Zionist but instead a lifelong citizen of the land of Goethe and Schiller, even after that land had ceased to exist except as an idea.” B and D both reach this independently and both add the post-Auschwitz complication Sophie’s introductions leave implicit — B: “Israel is the only answer the Jews could make to Auschwitz… I hold both of these at once”; D names the Brit-Schalom / Ichud position (Buber, Magnes, Ernst Simon) by name, which the German scholarship gave it and which neither Sophie nor A specifies.

Where the readings diverge on the person

The divergences are matters of emphasis and idiom, not of a different person:

Headline for §1: the person is highly recoverable. On who Tergit was, the two independently built reads land essentially where Sophie lands, differing mainly in how much named German institutional and biographical detail they carry — more in B and D, because their evidence base holds more of it.

2. Tergit-the-writer

Where all four converge on the project and the voice

If the person is recoverable, the writer is recoverable to an almost uncanny degree — the independently built reads reconstruct Sophie’s account of the prose from the prose itself.

A novel about people, not ideas; diagnosis, not prophecy. This is the through-thesis of every reading. Sophie: “Rather than a novel of ideas, Tergit wrote a novel about people… ‘I’m interested in people, men and women, and I’m interested in the ways their behavior changes over the decades. I delight in both their folly and their wisdom.’” A reproduces that sentence almost verbatim: “People, not ideas. I am interested in men and women and in how their behavior changes over the decades, and I delight in both their folly and their wisdom.” B, from the German side: “Diagnosis, not prophecy. I give the disease of the time; I do not pretend to foretell the catastrophe.” D: “ideas only matter insofar as they matter to her characters” (its step-2 gloss on Boa), and in the persona “I work in the social Typus, not in the inner movement.”

The object as historiography. All four read the menus, dresses, prices, and furniture not as decoration but as analysis. Sophie: “Her party scenes, descriptions of menu planning, and discussions of dress fabrics are not frivolous details but sophisticated analyses of class and society, similar to Edith Wharton’s dissections of Edwardian society.” A: “The little detail as the whole of society. … What a woman chooses to wear to a party… this is class, money, and the changing place of women, told without a single grand theory.” B: “the silver egg-cups on the lace cloth while the world burns… The documented thing is the time. Do not abridge it.” D: “Objects carry historical density. A Crown-Gasglocke in a Tiergarten dining-room in 1898 carries the whole of Wilhelmine bourgeois materiality.” Both B and D independently fix on the very objects Boa singled out (the Crown fixture, the toilet-roll holder with lion and unicorn) as the type-case of object-as-time-marker.

The chorus: dialogue carries the meaning; the narrator withholds judgment. Sophie: “Dialogue tends to be rapid-fire, with few descriptions of character reactions… she sticks to ‘he/she said,’ and the dialogue itself.” A: “Dialogue, rapid-fire, tagged bare. I stick to ‘he said,’ ‘she said’… I do not decorate speech with reaction-verbs and adverbs.” B: “The chorus. My books are largely dialogue… with almost no connecting, commenting, judging narrator between them… The evaluating ‘I’ of the storyteller is nearly gone.” D: “Conversation as primary action. People are who they say… The narrator’s voice in my novels does not own the judgment.” This is the same formal observation reached four times from three different evidence bases.

The flat death-sentence; laconic repetition. All four name the technique by which the worst news arrives in the plainest clause. A: “The flat death-sentence. … ‘Heinz is no longer alive.’ Let it land cold.” B: “A death is told in a flat half-line dropped between pleasantries — die sind ermordet worden.” D: “The elegiac final cadence. ‘Er starb 1943. Aber das erfuhr niemand.’ … the historical fact dropped into the sentence like a stone into water.” Sophie locates the same restraint in the prose and warns it must survive translation. A and B both also name the laconic repetition (“Taxes and taxes and taxes”; “Unchanged, unchanged” / “So war es? So war es.”) that B distilled directly from the primary corpus.

Comedy over an abyss. Olden’s phrase Lachen im Elend des Zugrundegehens anchors three of the four. Sophie (in the Käsebier intro): “Despite its ebullience and sprightly repartee, Käsebier is ultimately tragic… the spirit of Minerva atop the crumbling newspaper headquarters.” A: “Comedy over an abyss. The surface is buoyant, witty, sprightly; the undertow is tragedy… the spirit of Minerva on the roof of the crumbling building.” B: “Laughter in the misery of going-under. … Lachen im Elend des Zugrundegehens. … A book may be resigned without being resigning.” D names it through Olden’s review too. (The Minerva image is specifically Sophie’s, and A — built from Sophie — carries it; B and D, without Sophie, do not have Minerva but have the same doubleness from Olden.)

The cardinal law of Effingers: not written from the end. This is the single most important shared reading, and the convergence of the German-built reads with Sophie here is the heart of the study. All four insist the novel is not written backward from the catastrophe; the characters do not know the future; the antisemitism is sheet-lightning on the horizon, not thunder in the room.

B reaches this via Boa and von der Lühe (“the novel was not born out of the catastrophe”); D reaches it via the same Boa essay; A reaches it via Sophie, who reaches it via Samuel Moyn. Three routes, one reading. The phrasing “warns, does not predict” appears independently in both A and B. That the self-built German reads land on the exact tonal law Sophie articulates is the strongest single piece of evidence that the project of Effingers is recoverable from Tergit’s own work plus its German context, with no help from Sophie.

Neue Sachlichkeit warmed by appetite. All four place Tergit in New Objectivity and then qualify it with warmth. Sophie: New Objectivity’s “crisp, unadorned prose,” but with “singular humanity and compassion.” B: “This is Neue Sachlichkeit, yes… but warmed by a great appetite for the whole social world, the appetite of Balzac and of Fontane.” D: “neusachliche… the Type against the Bild… warmed.” A: New Objectivity “material and social description over interior psychology.”

Where the readings diverge on the writer — the lineage question

Here is the one substantive divergence in the reading of the writer, and it is the seam the whole study is built to expose: where do they locate Tergit’s literary kin?

So on the writer, the readings split cleanly along the evidence base: German evidence yields a German-lineage Tergit (Fontane’s heir); Sophie’s evidence yields an anglophone-kin Tergit (Wharton’s cousin); A inherits Sophie’s anglophone kin. The crucial wrinkle — that D, despite a German evidence base, also names Wharton/Powell/Mitford — is the subject of §4, where it turns out the two are doing very different things with the same names.

3. The work of construction — do the self-built reads find Sophie’s Tergit?

A had Sophie’s writing; B and D did not. So the test is direct: building only from Tergit’s own work plus context, do B and D arrive at the Tergit Sophie arrived at, or at a different author?

The answer is that they arrive at substantially the same Tergit — and the overlap is large enough, and specific enough, that it cannot be coincidence. What B and D recover independently is not just the biographical outline (that much is in any encyclopedia) but the interpretive core that is Sophie’s distinctive contribution:

  1. The anti-determinist law of Effingers — “warns, does not predict,” the people not living under the sign of apocalypse. B and D both state it; Sophie states it; A inherits it. This is not a fact to be looked up; it is a reading, and three evidence bases produce it.
  2. The novel-about-people-not-ideas thesis, down to shared phrasing about delighting in folly and wisdom (Sophie and A share the quotation because it is Tergit’s; B and D reach the same thesis through Boa’s “ideas only matter insofar as they matter to her characters”).
  3. The object-as-class-analysis reading — menus and dresses as historiography, not decoration.
  4. The withholding narrator / dialogue-chorus as the formal carrier of meaning.
  5. The even-handed anti-totalitarian, anti-generalization ethic as the moral spine.
  6. The Bildung-as-Jewishness, rooted-German-Jew, anti-Zionist-with-late-ambivalence religious- political profile.

The most economical demonstration is B’s twelve-point “voice signature,” distilled in step 1 purely from Tergit’s prose, before B had read a single word of secondary material, let alone Sophie: terse laconic declaratives; sparing commas and fronted predicates; laconic repetition as rhythm; catalogues as documentary argument; embedded speech fixed to age and class; “the catastrophe stated plainly, in a subordinate clause… Restraint is the carrier of grief”; money and objects as the ground of character and morality; the small detail standing for the whole; comedy that earns its gravity, “never solemn.” Set that list beside Sophie’s “Translating Effingers” section — direct, unadorned prose; rapid-fire dialogue with few reaction descriptions; the danger that the sparseness goes “stark” in English; the montage of headlines and song-scraps; honorifics and objects handled just so — and they are describing the same writer in the same terms. Sophie heard this voice and reached for an English to keep it; B heard the identical voice and reached for the identical description. The overlap of an independently built read with Sophie’s is the measure of how much of Tergit is in Tergit.

Where the divergence exposes what each evidence base does and does not contain. The places B and D differ from Sophie are exactly the places where their evidence base carries something hers does not, or lacks something hers supplies:

A subtle asymmetry is worth naming. Because A read Tergit’s primary corpus and Sophie, A’s convergence with Sophie is overdetermined: many of A’s “Sophie-derived” claims were already in A’s step-1 notes from the primary text before A opened Sophie. A’s step-1 carry-forward already had the recording-band method, the laconic style signatures, Eggebrecht’s “gedämpft, delikat, still,” the even-handed horror at both totalitarianisms (including the “pure orphans” line, which is Tergit’s own), the anti-Zionism, and “Ich bin Berlinerin.” So A is not a clean test of “what Sophie adds”: A is a reader who found much of Sophie’s Tergit in the corpus first and then had it confirmed. The cleaner test of recoverability is B (corpus + German context, never Sophie), and B passes it — B finds Sophie’s Tergit. The cleaner test of what specifically Sophie adds is the gap between B and Sophie, which is the anglophone re-homing.

4. What the anglophone corpus changed — B vs D (the central probe)

B and D are the cleanest comparison in the study: both self-built, both excluded Sophie, both worked the same German materials. The single difference is D’s added shelf — Wharton, Powell, Mitford, Isherwood. The question is whether that shelf changed D’s reading of Tergit (who she is, what her project is) or mostly furnished D with English voice-models layered over a reading that is otherwise B-like.

The finding is unambiguous: the shelf gave D a translation apparatus, not a different Tergit. D’s reading of who Tergit is and what Effingers is remains B-like in every load-bearing respect; the anglophone material is concentrated in a separate “how to render her in English” compartment, with two narrow exceptions where it lightly tints — and ratifies — a reading D already held on German grounds.

The structural evidence: D quarantines the anglophone shelf

D’s own step-2 notes are organized into labeled parts: A — BIO, B — PERIOD, C — PEERS, D — TRADITION, E — ANGLOPHONE (the English voice-models for Step 4), F — INTEGRATIVE. The anglophone authors appear only in Part E, which D titles, in its own words, “the English voice-models for Step 4.” Parts A through D — everything that constitutes the reading of who Tergit is, where she comes from, and what her project is — are built entirely from German sources. The tradition section (Part D) names Tergit’s masters as Fontane, Mann, Roth, Heinrich Mann — not one anglophone author. And each anglophone author in Part E ends with an explicitly craft-keyed line: “Translation principle from Wharton,” “Translation principle from Isherwood,” “Translation principle from Powell,” “Translation principle from Mitford.” The shelf is, structurally and by D’s own labeling, a drawer marked how to render her in English.

This carries straight into the persona. In D’s before-novel persona the four anglophone authors appear under one heading — “The English I will write Chapter 25 in” — and nowhere else. The sections that construct the self (Who I am; The Berlin I am from; How I was formed; The break; The writing self; The political-ethical position) are pure German lineage. When D summarizes its own identity in the integrative Part F, it reaches only for German coordinates: “a writer of the Chronistin-and-Kritikerin position… the Type against the Bild; conversation-as-action; Fontane-tradition; the jüdische Familienroman as response to Buddenbrooks.” No anglophone author appears in D’s statement of who Tergit is.

B and D converge despite the different corpora — the reading is corpus-robust

Everything in §1 and §2 that B and D share, they share because the German evidence base produces it, and D’s anglophone shelf neither added nor altered any of it. Side by side, B and D give the same writing-self:

B (German corpus) D (German corpus + anglophone shelf)
Master “I take Fontane’s Berlin and his dialogue… and I cut away his leisurely chatter to the bone” “Fontane gave me my mode: Plauderei als Erkenntnis… I learned to write under him before anyone”
Counter-text “I write the family saga that is Buddenbrooks’ subject… without the pathos of artful periodic sentences” “with Buddenbrooks as my deliberate counter-text and without the Mannian high-style”
The cool vs. the hot “the German Untertan whom Heinrich Mann lashed with hot contempt, and I render him cool” “the difference between Fontane and Heinrich Mann, and I am on the Fontane side”
Final cadence “A death is told in a flat half-line… die sind ermordet worden” “The elegiac final cadence… Er starb 1943. Aber das erfuhr niemand. … the mood is Roth’s, not Mann’s”
Method “The chorus… almost no connecting, commenting, judging narrator” “Conversation as primary action… The narrative interpolations are short… sparing of judgment”
Characterization type over psychology; the chill “in the juxtaposition, never in the adjective” “Type, not psychology… not from interior monologue”

This is the same Tergit, twice — and it is B’s Tergit, the Fontane-line chronicler, reproduced by D on the identical German grounds. The added shelf left the reading of the author untouched. Where the reading of Tergit is concerned, the corpus is robust: anglophone novels in, the same German-lineage Tergit out.

Where D bears a specifically anglophone stamp that B lacks

The stamp is real, but it lives in the craft register, and it is of two kinds.

(a) Pure voice-modelling, in the drawer. D assigns each scene-type an English register-source: “Isherwood-clear in narration… Wharton-periodic in the Tiergarten interiors… Mitford-quick in dialogue… Powell-deadpan where the scene is mercantile or modern-urban.” This is a register-toolkit B simply does not have. B, facing the same English problem, has no anglophone authors to attune to; B’s entire instruction for the English is negative and German-derived — honor the plainness, keep the bare “said,” do not pad, do not “modernize.” So the difference the shelf makes is that D arrives at the translation desk with named English ears for each register, and B arrives with only Tergit’s German signatures and the injunction not to betray them. This is a difference in translation-craft equipment, not in the understanding of Tergit.

(b) Two kinship claims where the shelf lightly tints the reading — and ratifies it. In two places D lets an anglophone author cross from the drawer into the reading itself:

But two things keep even these from amounting to a recasting of Tergit. First, they confirm a reading D already had on German grounds — the Sachlichkeit documentary eye (Isherwood) and the bourgeois-interior class-irony (Wharton) are things D had already established from Neue Sachlichkeit, Fontane, and Boa; the anglophone authors give them an English mirror, not a new content. Second, D keeps the kin at arm’s length on the actual writing-self, exactly as it keeps Mann and Heinrich Mann at arm’s length: “The Wharton sentence is longer and more periodic than mine… she has a Henry-James-Schule I do not have”; Powell’s mature manner is “not my model”; even Mitford, “deeply mine” in quickness, is “upper-class English where I am bourgeois-Jewish-Berliner.” D reads Tergit beside Wharton and Isherwood, not as one of them.

Does D read Tergit through Wharton/Isherwood, or keep them in a separate drawer?

Overwhelmingly the latter, with a faint dose of the former. The decisive contrast is with Sophie, who genuinely reads Tergit through the anglophone novelists: for Sophie the recognition of kinship with Wharton/Powell/Mitford is a claim about where Tergit belongs — “more at home within Tergit’s host country and its literary traditions” — a re-homing of the author out of the German context “in which it initially found little resonance.” D never re-homes Tergit. D’s Tergit belongs to Fontane and the Weltbühne and the jüdische Familienroman; the English authors are how a German writer is carried into English. Same names (mostly), opposite function: for Sophie they mark Tergit’s literary home; for D they are the vehicle out of it.

The convergence-on-the-same-novelists, and an important caveat

It is striking that D, with no access to Sophie, names three of Sophie’s four anglophone models (Wharton, Powell, Mitford). But this is not independent rediscovery: the experiment supplied D with exactly those novelists (Wharton, Powell, Mitford, Isherwood) as its added corpus. D hitting Sophie’s trio is therefore largely an artifact of the seeding, not a sign that Tergit “naturally” summons those three names. The genuinely D-driven choice is Isherwood, the one author not in Sophie’s set, which D elevates to “the single closest English voice-model” on the same-city/same-years logic — a logic Sophie, focused on the novel-of-manners genre rather than the Berlin subject, does not use. So the convergence on Wharton/Powell/Mitford tells us less than it appears to; what it does show is that when handed the anglophone novel-of-manners shelf, D finds it apt for Tergit and deploys it much as Sophie does at the level of register — which independently corroborates Sophie’s craft intuition that these are serviceable English ears for this German prose.

Bottom line for §4. D’s anglophone corpus changed D’s translation toolkit, not D’s reading of Tergit. The reading of who Tergit is and what Effingers is stayed B-like and German-lineaged. The shelf supplied a set of register-models in a clearly marked “how to render her in English” drawer, plus two kinship touches (Isherwood’s camera, Wharton’s opera-box) that decorate and confirm — rather than redirect — a reading D already held. The probe’s result: the reading of an author is robust to this corpus change; what the corpus change buys is voice apparatus for the downstream task.

5. The revisions — B and D before and after the novel

Both B and D revised after reading Effingers in full. D’s revision is visibly larger than B’s. But size and depth run opposite to each other here: B’s smaller revision changes the reading; D’s larger revision mostly re-affirms the reading and adds craft.

B’s revision: small in extent, but it moves the reading

B’s after-novel persona is identical to the before-novel persona except for one inserted paragraph — a new “second law” of the book:

“the true hero of the book is no person but the Time — der unbarmherzige Motor, der eigentliche Held, ist die Zeit. Since 1914 we have not lived our lives; we have been lived — no longer master and lord of our own fate. Paul believes he is doing the shoving, and he is being shoved… the answer is the Time. So I neither punish my people with their fates nor let them master them: the current takes them, and I hold the reader’s eye on the swimmer, never on the flood.”

This is interpretive, characterological, philosophical — a genuine development of the reading of the novel. Before the novel, B’s Tergit insisted on diagnosis-not-prophecy and the withholding narrator; after it, B’s Tergit adds a metaphysics of historical determinism (the individual lived by the Time since 1914) that was not in the before-persona. It is the one place in either German read where the reading visibly grows from contact with the whole book.

It also opens a small but real tension with Sophie and A. Sophie makes much of Tergit’s decision to reject the title Der Ewige Strom precisely so the characters would not seem “ineluctably swept up by the movements of history itself”; A turns this into a vow — “I will not have my people swept along by History like driftwood. They made choices… individual choices matter, even under adverse circumstances.” B-after’s “we are lived… the current takes them” tilts the emphasis the other way, toward the flood. B is careful to hold both (“eye on the swimmer, never on the flood”), but the center of gravity of B’s post-novel reading sits closer to determinism than Sophie’s and A’s choice-centered reading does. Reading the whole novel pushed B toward the Goethe-epigraph current that Sophie reports Tergit pushing away from. This is the most interesting divergence the revisions produce — and it is invisible without comparing the after-personas to Sophie.

D’s revision: large in extent, but mostly re-affirmation plus craft

D appends a four-point “Addendum after Step 3” that is longer than B’s insertion — but it opens by declaring it is not a change of reading: “The persona above holds. The reading did not change the self; it sharpened four things I now want to write down, because they will govern the chapter-25 English.” The four points are:

  1. Form is the signature — chapter 25 is the first of four polyphonic “Saturday-chapters” across seventy years (1887, 1913, 1930, 1948) with a recurring spring refrain, “Was für ein Frühlingstag… Was für eine Süße”; the English must be redeployable four times. (Craft / form.)
  2. The keyword is Süße / sweetness — keep the word, do not paraphrase. (Craft / lexis.)
  3. The frame is two letters (ch. 1 “Ein Brief” / ch. 151 “Ein Brief”); model the embedded letters on Daniel Deronda or The Wings of the Dove. (Craft / form.)
  4. Lightness is the law, confirmed across 900 pages — “the narrator does not raise her voice… The English must not raise its voice either.” (Re-affirmation of the existing tonal reading.)

Three of the four are translation-craft discoveries (the formal architecture of the refrain, the lexical keyword, the letter-form register), and the fourth is an explicit confirmation of the “not written from the end / do not raise the voice” reading D already held. D closes: “The four English voices I named in Step 2… are confirmed. Of the four, Wharton is the closest single model for chapter 25 specifically.” And the final line: “The persona stands as written. The reading has tightened the grip, not changed the hand.”

Two of D’s “discoveries” were in fact already in D’s pre-novel material. D’s step-2 notes (from Boa) already recorded “First letter, 1878… Last letter, March 1942… Epilogue 1948” — so point 3’s letter-frame is repackaged, not newly found. And point 4’s tonal law is the Boa/von der Lühe “not born out of the catastrophe” reading D held before the novel. The genuinely new products of D’s reading the whole book are point 1 (the four-Saturday refrain architecture) and point 2 (the keyword Süße) — both of them formal/lexical craft, the kind of thing only end-to-end reading surfaces, and both oriented entirely toward the chapter-25 English rather than toward who Tergit is.

D’s reach past its own corpus is worth noting: for the letter-register it invokes Daniel Deronda and The Wings of the Dove (Eliot and James), neither of which is in D’s given anglophone shelf (Wharton/Powell/Mitford/Isherwood). So once translating, D’s anglophone frame is generative — it expands to pull in further English models — but, again, only in the craft register.

Comparing the kind and size of the two revisions

One contextual cause for D’s revision being craft-weighted: D had deferred reading Effingers to step 3, so step 3 was D’s first contact with the book — yet D still treats the persona as standing, because D’s pre-novel reading (built from Boa’s close map of the novel plus the rest of the corpus) had already anticipated the substance. The whole-novel reading mostly filled in form and lexis. B, which had also worked from Boa, used its revision differently — to push past description into a thesis about the engine of the book.

6. A’s relationship to its source

A was built from Sophie. The question is whether A reproduces Sophie’s reading faithfully, or reorganizes, adds to, or underweights it — and whether any divergence is interesting.

A is a high-fidelity transposition of Sophie into the first person. Nearly every load-bearing element of A’s persona is traceable to a specific passage in Sophie, often with the wording carried across intact:

A’s own step-2 notes are candid about this: A identifies Sophie as “my English translator” and treats her essay as “the already-built English voice of my prose… continuity is itself a value. I write from them; I keep the last call.” So A understands itself as inheriting and consolidating a translator’s established reading.

What A adds — and it is the interesting divergence. A’s one substantive move beyond Sophie is a framing conceit: A is not Sophie-translating-Tergit but Tergit-translating-herself. A flags this honestly in its notes — “One tension to keep honest: I am Tergit translating herself, not Duvernoy translating Tergit — a subtly different stance.” From that conceit A generates material that is nowhere in Sophie: the claim of final authority over the text — “I am the author, not a faithful servant of the author. I write from that established voice, but I keep the last call… The book is not a museum of my style; it is the thing itself.” This is A reasoning as Tergit about Tergit’s own prerogative to override a translator’s instinct — a layer the source could not contain, because the source is the translator, not the author. It is the most genuinely additive thing in A, and it is a direct product of the experiment’s premise rather than of Sophie’s text.

Where A amplifies rather than reproduces. A is more aphoristic and sententious than Sophie. Some of A’s maxims compress things Sophie says discursively into epigram — “I say these things flatly because that is how the worst things are true”; “the figures are where the morality is”; “once lying is normal, truth itself becomes unprovable, and that is the abyss.” Several of these draw on Tergit’s own memoir-voice (which A had read in step 1) as much as on Sophie — for instance “The mob lays hands soonest on the gentle. The spirit is always quiet” renders Tergit’s own “der Mob vergreift sich am liebsten an den Zarten” / “der Geist ist immer leise,” which Sophie channels through Waldemar’s “those spoiled goods.” So A is not only transposing Sophie; it is fusing Sophie’s framing with the primary-corpus voice A had already absorbed, and sharpening both into maxim. The effect is a persona that sounds more like an essayist-Tergit delivering verdicts than Sophie’s more historian-like prose.

What A underweights. Relative to Sophie’s introductions, A compresses the dense historical- contextual apparatus — the Berlin urban history, the 1879–81 antisemitism debate, the Treitschke / Lazarus exchange, the demographic and architectural detail, the publishing-history of Effingers. A keeps the interpretive payload of all this (Bildung as Jewishness; the paradox that emancipation both integrated and marked out; England as the keeper of the failed liberal tradition) but sheds the scholarly scaffolding — appropriately, since A is writing a “self,” not an introduction. This is a reorganization toward voice, not a distortion of content.

Net for §6. A reproduces Sophie’s reading faithfully and at the level of specific wording; it reorganizes the material from third-person scholarship into first-person self-portrait; it underweights Sophie’s historical apparatus while preserving its interpretive conclusions; and it adds exactly one new thing — the author-translating-herself stance, with its claim to the “last call” — which is the experiment’s conceit made into character. The divergence is interesting precisely because it is the one place A cannot be merely echoing Sophie: it is A reasoning from a position (the author’s) that the source (the translator’s) does not occupy.

7. Patterns across the four

Drawing the threads together, several patterns recur across the readings.

Pattern 1 — The person and the project are recoverable; the literary home is not. The biographical-temperamental Tergit and the formal-tonal Tergit (chronicler not prophet, dialogue- chorus, object-as-history, restraint as the carrier of grief, anti-generalization) are reconstructed independently from three evidence bases and largely agree. What does not survive the loss of Sophie’s evidence is the placement of Tergit in a literary tradition: German evidence yields Fontane’s heir (B, D); only Sophie’s England-situated reading yields Wharton’s cousin. The recoverable layer is what Tergit is and does; the evidence-dependent layer is whom she is to be read beside.

Pattern 2 — A corpus changes the toolkit before it changes the reading. The B/D probe shows the added anglophone shelf re-equipping the translator (register-models, kinship analogues, an expanding set of English ears) while leaving the reading of the author essentially fixed. The understanding of Tergit proved robust to a substantial corpus change; the downstream voice-apparatus did not. If one wanted to change a reader’s reading of Tergit, this study suggests adding novelists to imitate is not how — the German interpretive sources (Boa, Wagener, the Fontane scholarship) are what set the reading, and B and D share those.

Pattern 3 — Drawer vs. through-reading is the real axis, not which authors are named. The surface fact “D and Sophie both cite Wharton/Powell/Mitford” conceals the deeper difference: Sophie reads Tergit through the anglophone novel of manners (it is where Tergit belongs), whereas D keeps the same novelists in a how-to-render drawer (they are the vehicle out of a German home). Same names, opposite epistemic function. The seeding caveat matters here too — D was handed those three, so the overlap is partly designed in; D’s own pick (Isherwood) is governed by subject-kinship (same city, same years), a different logic from Sophie’s genre-kinship.

Pattern 4 — Depth of primary steeping tracks where the reading is thickest. A and B, which read the primary corpus most deeply, produce the most prose-anchored accounts of the voice (A’s death- sentence/montage catalogue; B’s twelve-point signature). D, which steeped least in the primary text and deferred Effingers, produces the thinnest primary-voice account and compensates with secondary framing and the register exercise — which is also why D’s post-novel revision is where its formal/ lexical observations (the Saturday refrain, Süße) finally land. The evidence each reader weighted shows up in where its reading is most and least concrete.

Pattern 5 — Revision size is a poor proxy for revision depth. D’s large addendum mostly re-affirms and tools-up; B’s one-paragraph insertion is the only place a reading genuinely moves (toward the Time-as-hero determinism, and, in doing so, slightly away from Sophie’s and A’s choice-centered emphasis). The volume of change and the substance of change came apart.

Pattern 6 — The two stances toward authority diverge with the source. A, built from a translator, foregrounds the author’s authority (the conceit of self-translation, the “last call”). B and D, built as the author from scratch, foreground the chronicler’s authority (the duty to the truth, the witness against erasure). Sophie, the actual translator, foregrounds the translator’s authority (the attunement, the thousand small choices, the re-homing). Each reading’s theory of its own authority tracks the seat its evidence put it in.

None of these patterns ranks the readings. They describe how four constructions of one author, built from very different evidence, converge on who she was and what she did, diverge on where she belongs and how to carry her into English, and reveal — in the clean B-vs-D contrast — that an added shelf of novels reshapes the apparatus of translation well before it reshapes the understanding of the author.

Back to→ Appendix