The first of three main parts. It covers the habits and rules that belong to hadith scholars across the board, whoever they are: how to study, how to teach, how to handle transmission, how to write a manuscript, and how a text legally passes from teacher to student.
Two portraits: the student and the finished scholar. Baqa'i's angle is that character training was treated as part of the method, not as an extra.
The student's code, which the scholars filed under adab (manners). Sincere intention first; share what you learn instead of hoarding it; go step by step and be patient; act on what you hear; and honour your teacher. Whole books were written on this, such as al-Khatib al-Baghdadi's al-Jami' li-Akhlaq al-Rawi.
The teacher's code. Keep your intention clean and beware the love of being in charge; let your conduct match what you teach; do not set yourself up to teach before you are qualified; stop teaching if you fear your memory is slipping; and point students to someone better than you when there is one.
Moves from the person to the process: where the care taken in transmission comes from, how the Companions actually did it, and how the scholars after them received and passed on material.
The scriptural roots. Baqa'i grounds careful transmission in four things: lying is forbidden, guessing is forbidden and you must verify, a report from an unreliable person must be checked, and a witness must be of good character. The point is that the caution is religious in origin, not a later invention.
What the Companions actually did after the Prophet died. Four working habits: narrate only as much as is needed, check a report even if it means travelling, measure a report against the Qur'an and the known Sunna, and hold back material ordinary listeners would misunderstand.
Two habits. Go slowly: two to five hadiths a sitting, so they were mastered. And collect wide, screen late, Abu Hatim al-Razi's taqmish then taftish: write down all you hear, weak narrators included, and sift at the point of use rather than the point of collection. The looseness is about what enters your notebook, not about what you will vouch for.
A practical chapter on the craft of the manuscript, built on the two sections of Ibn al-Salah's 'Ulum al-Hadith about writing hadith down and keeping the copy accurate. Baqa'i regroups those scattered notes by topic.
Good practice at the desk. Vocalise and dot only what could actually be misread, and take special care with people's names, since a name cannot be worked out from the sense. Write clearly. And write the full prayer on the Prophet every time he is named, never shrinking it to a letter.
Two marking habits. Undotted letters get their own sign so a reader knows the dots are missing on purpose, not by accident. And a small circle goes between every two hadiths; it is left empty until the copy is checked against the original, then a dot is added for each round of checking.
What to do when the copyist skipped something. There is a set way to draw a line up from the gap, bend it toward the margin, write the missing words there, and mark the end with sahh. Ibn al-Salah even explains why it goes upward and to the right: so a second slip later still has empty margin to use.
Two opposite marks, easy to confuse. Tashih (تصحيح) writes sahh over wording that is right in both transmission and sense but looks doubtful, to show it was checked, not overlooked. Tadbib (تضبيب) marks wording that really was transmitted that way but is broken in language or sense, or incomplete.
Three ways to cancel text that does not belong, and which is best. Striking through (darb) is preferred, because the cancelled words stay readable. Scraping with a knife (kasht) and rubbing out (mahw) are disliked, since what you destroy may turn out to be correct in another version.
How to record more than one version in one copy. Build the body of the book on a single version and never splice two together. Put the differences between the lines or in the margin, each tagged to the person it came from. Al-Yunini's copy of Sahih al-Bukhari is the famous example.
The shorthand. Frequently repeated phrases get abbreviated, but only where the short form is obvious. Note one symbol in particular: the letter ha (ح) marks the point where the writer switches from one chain to another for the same hadith. It is very common in Sahih Muslim.
The attendance record on a manuscript. The student writes, at the front or in the margin, who heard the book read, from whom, and on what date. Baqa'i stresses honesty here: name people in wording that cannot be twisted, and never leave a listener out for a bad motive.
The legal heart of transmission. There are exactly eight recognised ways a text can pass from teacher to student, and each has its own honest wording for saying so later. They are ranked, roughly strongest first.
Way 1, and the strongest. The teacher speaks and the student listens, whether dictated or not. This is mostly how the Companions took hadith from the Prophet. You report it with sami'tu ('I heard') or haddathana ('he narrated to us').
Way 2. The student reads the text back to the teacher, who listens and confirms it; most scholars call this 'ard. You report it with 'I read to so-and-so'. Scholars argued over whether this ranks below hearing or equal to it; Abu Hanifa and Malik are reported to rate it highly, and Baqa'i thinks the gap is small.
Way 3. The teacher gives permission to narrate a text without the student having heard or read it back, e.g. 'I permit you to narrate Sahih al-Bukhari from me'. Baqa'i explains it arose once reliable written copies existed. Scholars counted nine kinds, from a named person and named book down to open-ended grants.
Way 4. The teacher physically hands the student the book. It comes in two kinds: handing over together with permission to narrate, which is the stronger, and handing over with no permission attached. Baqa'i notes a permission attached to a book still does not make the student a master of its contents.
Way 5. The teacher writes the material out for the student, or has it written, whether the student is present or in another town. Again two kinds, with or without permission to narrate. Al-Bukhari argued for it from the Prophet's own letters sent to rulers.
Way 6, and now we are into the disputed ones. The teacher merely informs the student that this book is his own transmission, without telling him to pass it on. Some scholars allow narrating on that basis and some refuse. Important: even those who refuse still hold you must ACT on it if the chain is sound.
Way 7, and the weakest of the handover routes. A teacher wills his books to someone on his death or departure. Al-Qadi 'Iyad reports early permission to narrate on this basis, but Ibn al-Salah rejects it outright, calling it either a scholar's slip or a case really meant as wijada, the next way below.
Way 8. You simply find something in a scholar's own handwriting, with no contact between you at all. You must say so plainly: 'I found in the handwriting of so-and-so'. Al-Nawawi classes the chain as broken, though the wording gives it a trace of connection. They agreed you may not narrate by it; they argued over whether you may act on it.
Why the exact wording matters, and this is the payoff of the whole chapter. The phrase tells you how the narrator got the text, so it feeds straight into judging the hadith. Say haddathana for something you only received by permission and you have made it sound as if you heard it directly; Ibn Hajar names scholars who did.