You are answering one question:
What problems was the Regency of Algiers actually trying to solve between 1785 and 1816?
Hold on to the shape of that sentence, because everything else follows from it. Algiers is the subject. The American texts — Tyler's novel, Foss's captivity journal, Rowson's play, the 1795 treaty — are evidence, one category of source among several. They are not the object of study.
This matters because the obvious version of your project is the opposite one. The obvious version asks how did Americans imagine Algiers? That question has been answered for thirty years by Allison, Baepler, Marr, Peskin and Berman, and an Americanist with better archives will always answer it better than you. Ask it and you have written a competent article that a hundred other people could have written.
Your question is different in a way that is not cosmetic. Ask what the Regency was solving, and Tyler stops being the thing analysed and becomes one witness among many — sitting beside Jefferson's letters, the treaty, the French historiography and the Arabic sources. That demotion is the whole strategic move of your career.
Because the Algerian side of this history is documented almost entirely in French and Arabic, and the Americanists in this field read neither. That is not a moral failing on their part; it is an accident of training that has left the scholarship one-sided for three decades. You read both. You also have something no American assistant professor has: the capacity to run a supervision programme, which means twelve dissertations over five years covering ground you could never cover alone.
Be honest about the limit of that advantage. You are not sitting on unknown documents. Belhamissi and Merouche are secondary sources; reading them is not working the Regency archive, and Ottoman-Turkish palaeography is training you do not have. The accurate formulation is: I can connect scholarly worlds that rarely communicate. That is a synthesiser's advantage. It is genuine, it is publishable, and it is enough.
Five years out, the test is this. An American scholar hits a question sitting between early America and the Regency of Algiers. Do they write to you — not as a fact-checker on Algerian detail, but because they know you consistently illuminate that seam in ways they cannot?
The label to aim at is an Americanist of the Mediterranean. Not "Algerian Americanist," which is an institutional category with a low ceiling. But understand that recognition is conferred, not chosen: it follows from where the first articles land, not from what you call yourself.
The cost, which you should accept deliberately rather than discover later, is institutional homelessness — never fully belonging to Early American Literature or to Maghreb history. For a scholar on the US job market that would be fatal. You have a school, a post, and no job market. You can afford it.
One submitted article, by 15 July 2027.
Article 1: terminology. The English corpus works with captive, slave, bondage, redemption, ransom. The Regency's own vocabulary keeps three things apart that English runs together: asīr, a prisoner whose status is recoverable through ransom or exchange; ʿabd, chattel condition; ʿilj, the renegade or convert. The English terms collapse a distinction the Arabic keeps sharp.
That collapse is not an accident of translation. It is what enables the antislavery analogy that dominates the existing scholarship — Tyler equating Algiers with Carolina. If captive and slave name different institutions, then that equation is doing rhetorical work rather than descriptive work, and the line of argument running from Baepler onward needs restating.
That is your article. Everything in the year serves it.
Every publication must require at least two scholarly traditions, both load-bearing.
The test is subtractive and you can apply it before you write: if the Algerian side could be cut without damaging the claim, the article isn't yours yet. Ask why you are the person writing it.
There is a formal name for this, which is worth having because it converts a personal working principle into a recognised methodology with a literature behind it: connected history, or histoire croisée. Werner and Zimmermann's objection is that comparison requires stable pre-defined units and a fixed observer position, and historical material usually refuses both. Their demand is reflexivity about the categories the historian brings — because those categories usually come from one of the traditions being studied.
Take that demand personally. Your categories arrive partly from the American scholarship you have been reading. That is the drift this whole system exists to prevent, stated as a methodological principle rather than a discipline problem.
Gravitational drift. The default outcome is derivative American studies with an Algerian name attached. The pull is strong because the American material is easier to reach and the existing scholarship supplies its own questions. The evidence this is real: two independently generated research plans for this project, both sophisticated, neither containing a single French or Arabic title.
Best reader, not leading interpreter. Complete mastery of the corpus produces authority about facts. Only interpretation produces citation. You do not want to be cited as the man who checks Algerian details; you want to be unciteable-around. Note that this failure mode is in direct tension with "know the corpus completely before publishing" — which is why coverage must arrive through the work, not before it.
| Window | What happens |
|---|---|
| August (now) | Hunt Belhamissi's captives book — the critical path. Readex email due 20 Aug. Method reading begins (capped). |
| September–October | Merouche, gate core — 6 hours of targeted chapters. Grammont quarry pass. |
| November – 15 December | Belhamissi. Pearce's genre article at the end. |
| 15 Dec – 5 January | Write the five-page brief. Build the generic template. |
| 10 January | The gate passes. The corpus opens. |
| January–March | Foss, Tyler, Rowson — research sheets, read against the brief. |
| 25 March | The 800-word argument. Two gates applied. |
| April–July | Draft, revise, submit. Panzac and Shuval as background reading. |
This is the most important structural decision in the plan and it is worth being able to reconstruct the reasoning.
You cannot read the evidence before you hold the criterion. If you spend the autumn reading forty-nine hours of American texts about Algiers, you will absorb the American categories — despotism, piracy, slavery — as your default frame, and the counter-sources will arrive afterwards as corrections to a picture already set. Gravitational drift is not a character weakness; it is a reading-order problem.
Reverse the order and the Regency's categories become the default. The American texts then arrive as the strange object to be audited, rather than the frame you are thinking inside. That is what "put Algiers on the subject side of the sentence" means in practice, applied to your own head.
Put more simply: you can only see errors where you already hold a standard. An error-list built before you know the Regency catches exotica and internal contradictions. An error-list built after catches what actually matters.
The gate is two books, not four: Merouche on la course, Belhamissi on the captives. Panzac and Shuval are context, not foundation — they moved to background reading during the drafting months, and they must not be allowed to delay anything.
A gate has to be finite and defined in advance or it expands forever — which is the year-with-no-output failure wearing scholarly clothes. So the gate ends in a thing you can hold:
A five-page brief on how the Regency's captivity economy actually worked: ransom brokerage, the fiscal role of la course, the legal status of captives, the vocabulary, what problems the Regency was solving, and the generic template. Written before you open Foss.
When the brief exists, the gate has been passed. That is the only test.
And the brief is not bureaucracy — it is what makes the later reading critical instead of absorbent. With it in hand you are not asking how does Tyler represent Algiers; you are checking his account against a description you already hold. Without it you would read Tyler and simply not notice what isn't there. Omission is where the findings are, and only a positive description of the Regency exposes omission.
Merouche is on your disk. **Belhamissi's Les captifs algériens et l'Europe chrétienne (ENAL, Alger, 1988) is not, and it has never been digitised** — catalogued in Tunis, listed on Google Books, occasionally at French auction, absent from Archive.org, Cairn and OpenEdition. Since it is an Algiers imprint, your best odds are local: the Bibliothèque Nationale d'Algérie and the Algiers university libraries, then CRASC, then French secondhand, then interlibrary loan through UKM.
The risk is partly covered. You do hold Belhamissi's Marine et marins d'Alger trilogy, which gives you his framework, his glossary, and the treaty system in his own words — four of the brief's six sections. What it cannot give you is ransom brokerage and the legal status of captives. If nothing arrives by mid-September, write the brief with that section flagged as an acknowledged gap. Do not let a missing book cancel the year.
You are not "doing postcolonialism," and you are not doing pluralism either. Article 1 declares two methods, not eight. Methodological pluralism is something you earn after publications; announced at the outset it reads as not having decided, and a methodology that commits to nothing cannot produce a finding.
Declared in the article:
Architectural, not declared per article: connected history keeps both traditions load-bearing. New Historicism supplies the plane on which a novel, a pamphlet, a journal, a treaty and a Congressional debate can all sit as one corpus without a hierarchy of literary over documentary. Use it for the plane only — it has no external referent, it reads representations against other representations and stops, and your project requires a check outside the discourse.
Available but not declared: microhistory, book history, Mediterranean studies, reception theory, intellectual history. These arrive when a specific article needs them. Microhistory in particular is a strong future fit for Cathcart — a crossing figure documented from both sides, with Natalie Zemon Davis's Trickster Travels as the model. But microhistory needs a dense archive around one case. Cathcart may support it. Foss does not, and the label on a thin archive is anecdote with footnotes.
Cited to depart from: Orientalism. Not because Said is wrong, but because he does not describe this configuration. Three reasons, and you should be able to give them without notes: the power vector is inverted — in 1785–1816 Algiers is the stronger party, holding captives, setting terms, taking tribute from a republic that cannot answer militarily, whereas Said's archive is discourse produced by the dominant about the dominated. It preserves the object relation — it still asks how Americans imagined Algiers, which is the paradigm you are leaving, in a more sophisticated register. It is the crowded default in Algerian English departments, so adopting it makes you legible and generic in a single move.
Cite Said to mark the departure. Establishing why Orientalism does not fit is a stronger opening than any application of it, and it demonstrates command rather than avoidance.
Proportion: two paragraphs of methodology in the article. No more. The frame's job is to make the reading possible and tell reviewers where you stand. Everything beyond that is decoration, and decoration is what makes literature articles unreadable.
This section is the heart of the handbook. Read it twice.
An intention is a method you believe in. An instrument is a method you cannot skip, because something in front of you has a blank that must be filled.
Every scholar intends to read critically. Almost none can say, six months later, which specific claims they tested and how. The intention evaporates because nothing recorded it. So the two filters have been built into tools with fields you fill while reading — which means the method now happens whether or not you remember to be rigorous on a given morning.
Three steps, applied to every claim an American text makes about Algiers:
Here is the worked example, and it is the sentence to remember:
Foss's descriptions of the master's cruelty are what the captivity genre demanded and what served his purpose in publishing. Weak evidence, whatever the truth. Foss's incidental details — prices, rations, work-gang procedure, who did what in what order — are things he had no convention requiring and no motive to invent. Strong evidence. And they are precisely what the representation-focused scholarship has passed over for thirty years, because it was reading these books for their portrayals rather than their residue.
That last clause is your finding waiting to happen.
Where: /research/algerine/genre/ · Build it by: 5 January 2027, before the corpus opens.
What it is. A list of the elements the Barbary captivity narrative was expected to contain — the taking at sea, the sale, the description of the city, the master's cruelty, the temptation to convert, labour and the daily round, providential framing, the return, Islam described, the renegade. The form inherits from two lines: the Indian captivity narrative (Pearce 1947, Slotkin 1973) and the Puritan spiritual autobiography, which supplies the shape of affliction → endurance → deliverance.
Why it exists. Every one of those elements is a topos before it is a report. A historian reading Foss for facts has no principled way to separate what he saw from what the genre obliged him to write. You do — but only if the repertoire is written down before you start reading, which is why the deadline sits before the corpus opens.
How to work it. Read Pearce and Baepler's introduction, then tick each element as you confirm it belongs to the repertoire. Each entry carries a note on generic pressure: high pressure means low evidentiary value. Then, reading any corpus text, ask of each claim: did the form require this? If yes, it is weak. If no, it gains weight and goes to the ledger.
The inverse move, where the finding is. Set the template against Merouche's quantitative data on prizes, ransoms and the fiscal organisation of la course. Where a text reports something the genre did not require, and the counter-sources confirm it, you have testimony nobody in this scholarship has used.
Where: claims-ledger.xlsx in Downloads (since 11 Aug 2026; the old site page /research/algerine/claims/ is retired and signposts the move). The two filter fields and the computed verdict are columns F to H of the Claims sheet; the constraints register is its own sheet.
What it is, from zero. A running list. Each row is one thing an American text asserts about the Regency — an institution, a motive, a price, a procedure. You write the claim in one line, with the text and the page, and you assign it a status. Nothing more.
The rule that makes it work: on first reading, you log, you do not judge. Every claim enters as unverified. This is deliberate. You do not need the criterion to notice that Tyler says the Dey personally owns all captives; you only need it to decide whether that is true. Separating the noticing from the judging is what lets you start before the counter-sources are read — and it is why the ledger, unlike a note-taking habit, does not quietly become an opinion diary.
What you do later. When the counter-sources are read, you return to each row and adjudicate it: confirmed, refuted, or nuanced — recording the French page beside the American one. That pairing, row by row, is the two-traditions rule made into a data structure. It is also source criticism in its oldest and most respectable form: testing a text's assertions against an independent record.
The two filter fields. Each row also carries the two judgements: genre required it? (yes / no / unclear) and motive to falsify? (yes / no / unclear). The ledger computes the verdict for you. Both answered "no" returns TESTIMONY — survives both filters. Both "yes" returns convention + interest — weakest. Anything else is partial or still pending.
What comes out. A tally showing how many claims have become testimony and how many still await judgement, and an export to Bear or CSV carrying claim, page, domain, both filters, the verdict and the adjudicating source. That export is the evidence table of your article, assembled as a by-product of reading rather than reconstructed at the end from memory.
Where: /research/algerine/terms/
This one is not a filter — it is Article 1's evidence base, and it is historical semantics made fillable. As you read, log every instance of a status word — captive, slave, bondage, redemption, ransom — with its page and a note on register: was it narration, quoted speech, or a documentary list? It tallies captive-family against slave-family per text.
Why register matters more than raw counts: the claim is not that Americans never knew the difference. It is that English usage collapsed a distinction the Regency's vocabulary kept. If documentary passages in Foss prefer captive while suffering scenes prefer slave, the collapse is shown to be rhetorical rather than descriptive — inside a single eyewitness. That is the strongest English-side evidence Article 1 can have.
The tracker also logs Koselleck's asymmetrical counter-concepts: pairs like Christian / Turk, free / enslaved, civilised / barbarian, where one side names itself and names the other, with no reciprocity. Nobody in this scholarship has used that tool on this corpus.
For orientation, the raw counts already measured across the corpus: Foss uses slave-family words 124 times against captive-family 59. Stevens is the corpus maximum at 214. Cathcart — the insider, the man who worked inside the Dey's administration — is the one major text that is captive-dominant. That asymmetry is not decoration; it may be the article.
Everything lives at read.progress-dz.com. The front door has two wings, because you have two reading lives on two different clocks.
/ Home — two cards
├── /reading/ The Shelf — evening reading, the 21:15 slot
│ ├── /the-road/ (a book plan: ten nights, motifs, vocabulary review)
│ └── /whole-brain-child/
└── /research/ The Desk — the protected mornings
├── /research/algerine/ THE CORPUS ← your hub
│ ├── /regency/ THE GATE — Merouche, Belhamissi, the brief
│ ├── /ximenez/ the §2/§3 patch, in measured sessions
│ ├── /text-01/ … /text-11/ one study page per corpus text
│ ├── /passages/ THE PILE — every passage the slips capture
│ ├── /institutions/ the institution index, side-aware gauges
│ ├── /misattribution/ the misattribution register
│ ├── /terms/ the terminology tracker, Article 1's evidence base
│ ├── /genre/ FILTER ONE — the generic template
│ ├── /claims/ RETIRED 11 Aug 2026 — signposts to the spreadsheet
│ └── /handbook/ this document
└── /research/writing/ Belcher, resequenced — the writing programme
How you move (since 11 Aug 2026). Every page carries the same shell. A top bar shows a breadcrumb (Reading Room › Desk › Algerine › the page), so you always know where you are and every ancestor is one tap back. The ☰ button opens a drawer listing every page on the site, grouped by wing, with your current page highlighted; on a wide screen the drawer stays pinned as a sidebar. On long pages a row of chips under the bar jumps to that page's own sections: on a study page, ✂️ Capture · Vocabulary · Sessions · Quiz; on this handbook, one chip per numbered section. You are not expected to remember where anything lives. Open the drawer: the drawer knows.
Start at /research/algerine/. That page is the hub, and it shows four things at a glance: how many texts you have read, how many of the 68 reading sessions are done, the next deadline, and the instrument — the two-tradition gauge, currently 11 American to 5 Algerian. That gauge is a conscience, not a statistic. If the Algerian number has not moved by the time you draft the argument, the gate has failed and you will know before writing eight thousand words.
Below the gauge sit the eleven corpus texts in order of printing, each expandable to show its imprint, its reading sessions, and a button into its study page. Below them, the four French counter-sources and the dated road to Article 1.
Each text's study page follows the same three-phase shape: Before you read (working vocabulary verified against the actual text, plus one instrumental lens), Read (the sessions, which share their state with the hub — tick in either place), and After you read (retrieval questions with explanations, a harvest checklist tied to the roadmap, and a Bear button that opens the nine-field research sheet pre-filled).
Five tools, five jobs, and keeping them separate is what stops the system collapsing into one undifferentiated pile:
claims-ledger.xlsx, in Downloads) holds Filter Two's record: the claims with both filter columns and the computed verdict, and the constraints register (C01 to C61). It left the site on 11 Aug 2026 because adjudication is a matching job, and a spreadsheet does joins natively.One warning worth carrying: ~/Downloads/things_algerine_year1.py is archived — do not run it. It creates the Things project from scratch by opening a URL, so running it again duplicates everything. The live project has diverged from it far beyond recovery.
You will meet, most mornings, a passage that feels important. This section is what to do in the ten minutes after that feeling, and it exists because of two model passages from Merouche that were worked through in August 2026. Everything here is encoded in the passage slip on the gate page and the corpus study pages; this is the reasoning behind the form.
Ask them in order. The slip's fields are these questions wearing input boxes.
at capture — paraphrase smuggles your own categories in, and you will want the exact wording when you cite.
Merouche spans 1518–1830 and much of it describes a system that no longer existed when Foss arrived. A passage without a date is not evidence.
institution index assembles itself from these tags, and institutions are where the American imagination and the Algerian tradition intersect.
author's own English is not Algerian reality. Getting this wrong once collapses Article 1.
This question produces arguments.
with no destination is a note; a slip with a destination is a deposit.
Passages are not equal, and the test for rank is Claude's carry-forward rule: does this constrain what an American text can credibly claim? Facts are useful; constraints are arguments.
galleys were mostly gone by 1785: a fact about a dead technology needs a bridge before it can speak for your period.
chosen for lightness and speed; the bench mixed deliberately against mutiny.* Mechanisms travel better than facts because the problem they solve persists.
Algiers for everything; penalties were graduated and reached the raïs.* Set against the vocabulary of lawless piracy, this is the structural misdescription, evidenced.
Tripoli to escape punishment.* You can only flee a jurisdiction that exists: conduct that cannot be staged for the record, proving the penalties were real. Nobody had a motive to fabricate it. The highest grade there is.
The pile page counts your constraints and proofs separately from your facts. If the pile grows but that number does not, you are collecting a scrapbook, not an argument.
And above the ladder sits disconfirmation. The passages that damage your hypothesis are worth more than the ones that support it, because they are the only ones that can improve it. The model case, 7 August 2026: the fleet-asymmetry figures at Merouche II p.163 (Algiers's whole fleet 562–1,000 guns; the English navy alone ~10,000 by 1697, and Merouche conceding Western superiority "ne peut être contestée") landed on the Administered Course hypothesis, whose arc asked depoliticisation to cause the naval vulnerability. The gap predated the qubtân's reduction by decades — institutional arrangement cannot carry that causal weight. The hypothesis was demoted the same day from caused to failed to close, and what replaced the demoted step was stronger: if Algiers was outmatched tenfold in the 1680s and still extracting tribute a century later, the course was never primarily a military instrument — it was leverage, a weak state monetising nuisance value against far stronger ones. Which reframes 1785–1816: not a strong Algiers preying on a weak republic, but two comparatively weak naval powers, one of which had spent two centuries perfecting extraction of payment from stronger ones, and the other of which had just lost its protector.
The procedure when it happens to you: tick "cuts against my own hypothesis" on the slip; log a row in the ledger's ◆ hypothesis column; amend the hypothesis note in Bear the same day — a dated amendment, never a silent rewrite. A hypothesis note keeps its own history, exactly like the ledger.
One passage can carry two different kinds of value, and forcing it into one slip loses one of them. The model is the «vie à bord» passage:
du tout noir au tout rose, reflecting circumstance and authorial temperament both — is Method: source criticism of the captivity corpus performed by a Maghreb historian decades before you arrived. That is citable armour. When a reviewer calls the two-filter procedure undue scepticism toward Foss, you are not importing Anglophone suspicion; you are extending Merouche's own caution.
Institution evidence at the top of the ladder.
Two slips, thirty seconds apart, via the "another from the same page" button. Different registers, different destinations, same page number.
The gate's product is not facts about the Regency — it is the interrogation you will subject the American texts to. Because of Merouche you can now ask of Foss: is discipline institutional or individualised into "the cruel Turk"? Is administrative hierarchy visible at all? And the richer reframing: not is he lying, but how does a captive experience a highly regulated military institution as personal cruelty?
Bank these on the slip as you read. They resurface automatically on every corpus study page in January — the October you interrogating the January you, which is the whole architecture of this year in miniature. Many researchers collect facts. Good researchers collect questions.
Over the year, stop thinking in books and start thinking in institutions: the dey, the dīwān, the ṭā'ifa, the raïs, discipline at sea, ransom, captive status, the treaties. They are the stable reference points where the two traditions actually meet, which makes them the natural unit for the two-traditions rule. Every slip you tag lands on its institution's card at /research/algerine/institutions/, split into an Algerian column and an American column. An institution lit on both sides is a place an article can stand. When one matures, open its dossier in Bear from the card.
Two protected mornings a week. Here is the procedure, which is the same whichever phase you are in.
/research/algerine/regency/. Take the next unticked session in the gate core — that is Merouche chapters 10, 11 and 17–21, six hours in total, not the full seventeen and a half./research/algerine/text-07/ for Foss, say.- Every claim about the Regency → the ledger, status unverified, with both filter fields answered if you can. - Every status word with its page and register → the terminology tracker. - Dated events → the chronology.
The rhythm to protect: the instruments are filled during reading, not afterwards. A ledger row written from memory a week later is an opinion. A ledger row written with the book open is evidence.
And one thing that is not a task: the two protected mornings will not appear on their own. The threat to this project has never been motivation. It is 850 students and 55 staff.
If the argument passes both, write the article. If it fails either, you have saved yourself four months.
algerine-research-dossier.md, §1b for methodology) decides; the primer (critical-approaches-primer.md) describes each approach in full. Both are in Bear under #university/algerine and in sessions/2026-07-reading-plan/study/methodology/. This handbook is the map between them and the instruments.