Reading Lab

All questions and data files

A critical reading of the orders, regulations, memoranda and press releases: questions, not verdicts. Each question has its evidence (a link to its place in the text), its method, its confidence, its status and its conclusion. All questions in one file: all.csv

2Translation gap
27No Arabic text
10The final legislative rush
1References and amendments
21Date and text integrity
21The map of power in the orders’ language
1Figures versus claims
8Shifts in discourse
1Silences

Orders with no Arabic text in the archive, by month

Apr 03Jun 03Aug 031Oct 03Dec 03Feb 0418Apr 04611Jun 04

The pattern is clear in the archive: 2 of 64 before April 2004 against 25 of 35 after. Only the absence from the archive is established; the orders among them have an Arabic transcription from the Dorar al-Iraq legal wiki, taken from the Iraqi Official Gazette.

The final legislative rush

Apr 03410Jun 0336Aug 03152Oct 0324Dec 0358Feb 04514Apr 04714Jun 04

Orders by month. 21 orders in May–June 2004, a rate of 2.5 a week against 1.6 before; 17 have no Arabic text in the archive, and 10 address the period after the handover in their text.

Shortest Arabic texts relative to the English

Order 64 · Amendment to the Company Law No. 21 of 199729Order 68 · Ministerial Committee for National Security62Order 1 · De-Baathificaion of Iraqi Society68Order 12 · Trade Liberalization Policy72Order 71 · Local Governmental Powers75Order 40 · Bank Law78Order 35 · Re-Est Ablishment of the Council of Judges81Order 48 · Delegation of Authority Regarding an Iraqi Special Tribunal83

Arabic length as a percentage of the English (archive-wide median close to 100%). The difference may come from OCR, an incomplete file or another version of the order.

Orders most cited by “amending” orders

Order 38 · Reconstruction Levy2Order 2 · Dissolution of Entities1Order 39 · Foreign Investment1Order 45 · Non-Governmental Organizations1Order 47 · Amendment to Coalition Provisional Authority Order — NUMBER 1Order 54 · Trade Liberalization Policy 20041Order 37 · Tax Strategy for 20031Order 49 · Tax Strategy of 20041

Number of later orders with “Amend” in their title that cite the order.

Questions by confidence

Established by the text72Likely7Doubtful4Open question9

Established: the fact is written in the document. Likely: a strong automatic inference. Doubtful: may be an extraction artefact. Open: cannot be answered from the archive alone.

The map of power in the orders’ language (orders per month)

Apr 03May 03Jun 03Jul 03Aug 03Sep 03Oct 03Nov 03Dec 03Jan 04Feb 04Mar 04Apr 04May 04Jun 04
Administrator’s powers12221113221
Overriding Iraqi law1132
Immunity clauses1321113
Penalty clauses33114112133428
Consultation with Iraqis12624

Number of orders whose English text contains each kind of wording, month by month (of 99 orders with English text).

Shifts in discourse: share of documents using each term (%)

Apr 03May 03Jun 03Jul 03Aug 03Sep 03Oct 03Nov 03Dec 03Jan 04Feb 04Mar 04Apr 04May 04Jun 04
occupation / occupying power — الاحتلال2547812569921
liberation — التحرير735218128338108812
sovereignty — السيادة7228102813242425313946
terrorists — الإرهابيون882034242514111825262339
insurgents — المتمردون5645314311
former regime elements — فلول النظام14286343549
anti-Iraqi forces — القوى المعادية للعراق1121
foreign fighters — المقاتلون الأجانب121112252968

Percentage of the month’s documents (excluding contracts, budget, military training and miscellany) that use the term in English or Arabic. Briefings include journalists’ questions.

Figures stated in releases and briefings (strict patterns)

  1. 8 Aug 2003Briefing34,000 police · current — “ce force participated in repression by the state. Right now there are almost 34,000 Iraqi police on the streets. Of these, over 4,500, p”
  2. 24 Aug 2003Briefingover 40,000 police · current — “as fired from his job. He's a very brave man, and he is now heading up more than 40,000 police who are on duty, Iraqi police who are o”
  3. 26 Aug 2003Briefing40,000 police · current — “we basically had to create a new police force. We’ve now got 40,000 Iraqi police serving already all across the country, we inte”
  4. 31 Aug 2003Briefing300 MW · at liberation — “ecurity services did not trust you with mobile telephones. At the time of liberation, Iraq produced only 300 megawatts of electricity daily Today Iraq”
  5. 31 Aug 2003Briefingover 1 million barrels a day · current — “than 7.5 billion dinars in wages. Right now Iraqi workers are producing over 1 million barrels of oil per day. · All the money receiv”
  6. 3 Sep 2003Briefing37,000 police · current — “single policeman in the country. There was no rule of law. Today we have 37,000 police. We have 49 of the 151 prisons reopened”
  7. 19 Sep 2003Briefing34,000 police · current — “ts of justice? The Coalition has worked hard to ensure that 34,000 Iraqi police officers are now on the streets. And there are plans to increase the n”
  8. 23 Sep 2003Briefing40,000 police · current — “nd a communications network to link it all together. Already 40,000 police are on duty. Our plan will double this number in th”
  9. 9 Oct 2003Releaseover 40,000 police · current — “Iraqi Minister of Interior Nori Badran. Iraq currently has more than 40,000 police on duty throughout the country. The new facili”
  10. 17 Nov 2003Briefing2.2 million barrels a day · current — “: We continue to ramp up oil production and oil exports. We are producing approximately 2.2 million barrels a day, exporting just north of 1 millio”
  11. 19 Nov 2003Briefing2.2 million barrels a day · current — “ng domestic oil needs right now? MR. SENOR: Right now Iraq is producing approximately 2.2 million barrels per day. Iraq is exporting over 1 milli”
  12. 20 Nov 2003Briefing2.2 million barrels a day · current — “ng domestic oil needs right now? MR. SENOR: Right now Iraq is producing approximately 2.2 million barrels per day. Iraq is exporting over 1 milli”
  13. 1 Feb 2004Briefingover 60,000 police · current — “ived here, there were no Iraqi police on the streets, today there are over 60,000 Iraqi police on the streets. Certainly the presence”
  14. 2 Feb 2004Briefing8,000 police · current — “-- we just graduated a class of police from the academy -- we have almost 8,000 police. We'll have 10,000 by May, and 19,000 b”
  15. 10 Feb 2004Briefingover 70,000 police · current — “i police patrols. Here we are, some nine months later, and there are over 70,000 Iraqi police on the street. Could they improve? Shou”
  16. 1 May 2004Briefing2.5 million barrels a day · current — “cture is being rebuilt, with the Iraqi oil industry already producing about 2.5 million barrels per day. On the ground in Iraq, we hav”
  17. 20 May 2004Briefing300 MW · at liberation — “e exceeded prewar levels in terms of electrical generation. When we arrived here, Iraq was generating something like 300 megawatts of power a day. And”
  18. 24 May 2004Briefing4,200–4,400 MW · pre-war — “mmer. This summer we will be generating somewhere near the pre-war levels -- the prewar levels were 4,200 to 4,400 megawatts a day. We are now generat”
  19. 27 May 2004Briefingover 2 million barrels a day · current — “fic details to the Iraqi oil ministry -- but Iraq right now is producing well over 2 million barrels per day that -- or exceeded prewar leve”
  20. 9 Jun 2004Briefing2.4 million barrels a day · pre-war — “rget of the enemy. In recent months, production reached the prewar level of about 2.4 million barrels a day. Under Saddam, the revenue from t”
  21. 28 Jun 2004Release4400 MW · pre-war — “at was used to achieve the Coalitionís goal of reaching the pre-war level of 4400 MWís of electricity was called the ì4400 M”

Each figure with its sentence. Only like is compared with like: the same measure (serving, production, generation) and the same kind.

Periods with no order, regulation or memorandum (21 days or more)

14 October 2003 → 4 November 200321

Days between two consecutive dated numbered documents in the archive.

Translation gap 2

We compare each order’s Arabic and English texts: overall length (the archive-wide median is close to 1:1), and amounts written with thousands separators in English that do not appear in the Arabic, nor as millions or thousands. Almost all Arabic text is OCR, so most of these questions stay “doubtful” until checked against the original.

Download CSV

No Arabic text 27

Orders with no Arabic text in our archive (University of North Texas Libraries and the Internet Archive). We compare their share before and after April 2004. What is established is only the absence from the archive, not that no Arabic text was issued: the orders among them now have an Arabic transcription from the Dorar al-Iraq legal wiki, which states that it reproduces the Iraqi Official Gazette; the Gazette issues in which they were published have not been found.

Download CSV

The final legislative rush 10

Orders dated 1 May – 28 June 2004, and their weekly rate compared with earlier months; we search their English text for wording that extends their effect beyond the handover (remain in force, Iraqi Interim Government, after the transfer of authority…). The wording is established by the text; its legal meaning is a question.

Download CSV

References and amendments 1

From machine-read references: orders that cite a number missing from the archive, and orders that more than one later order amends in so many words (is amended, amendment to, replaced…). Reference reading is automatic and can misfire on OCR text.

Download CSV

Date and text integrity 21

We compare the official code date in the Arabic and English texts and across a document’s copies, and against the earliest preserved file date (including Internet Archive captures); we flag files whose names describe them as revised or amended, and orders without a readable dated code.

Download CSV

The map of power in the orders’ language 21

In each order’s English text we count: the Administrator’s powers (“the Administrator may/shall”), “notwithstanding” clauses that place the text above other law, immunity clauses, penalty clauses, and consultation with the Governing Council or the Iraqi government. The counts are established by the text; their meaning is a question.

Download CSV

Figures versus claims 1

From press releases and briefings we extract recurring national figures (megawatts, police numbers, million barrels of oil a day) and classify them automatically as a target, a baseline (“pre-war”, “at liberation”) or a current figure. We ask when the same baseline changes, or when the current figure falls between close statements. The figures are established by the text; whether they measure the same thing is a question.

Download CSV

Shifts in discourse 8

For each month we measure the share of documents that use a term (in English or Arabic) out of all that month’s documents, so that a rise in volume is not mistaken for a shift in vocabulary. First appearance in the archive is not first use in reality. Transcripts include journalists’ questions, so a term in a transcript is not always the CPA’s own word.

Download CSV

Silences 1

Periods with no dated order, regulation or memorandum in the archive (21 days or more), numbered documents that other documents cite but we do not hold, and topics present in press releases but rare in the orders. Absence from the archive does not mean non-existence.

Download CSV