diff --git a/hledger/hledger.m4.md b/hledger/hledger.m4.md
index c29d3ae35..6b8977c9c 100644
--- a/hledger/hledger.m4.md
+++ b/hledger/hledger.m4.md
@@ -3677,28 +3677,33 @@ you should declare the decimal mark explicitly with this rule, to avoid misparse
 
 ## CSV fields and hledger fields
 
-This can be confusing, so here's an overview. (It mentions some things we haven't covered yet.)
+This can be confusing, so let's start with an overview:
 
 - **CSV fields** are provided by your data file.
-  Their default name is their position in the CSV record, starting with 1.
-  You can also give them a readable name (with the `fields` rule).
+  They are named by their position in the CSV record, starting with 1.
+  You can also give them a readable name.
 
-- **hledger fields** are predefined.
-  They correspond to parts of a transaction's journal entry (mostly).
-  `date`, `description`, `account1`, `amount1`, `account2` are some of these.
+- **hledger fields** are predefined;
+  `date`, `description`, `account1`, `amount1`, `account2` are some of them.
+  They correspond to parts of a transaction's journal entry, mostly.
 
-- These are the only fields you'll work with; you can't make new ones in a rules file.
-  (If you need extra CSV fields, you can add them to the data in preprocessing, before running the rules.)
+- The CSV fields and hledger fields are the only fields you'll be working with;
+  you can't define new fields, or variables as in a programming language.
+  (But you could add extra CSV fields to the data in preprocessing, before running the rules.)
 
-- You'll be reading CSV fields. (They can't be written to.) They'll be on the right hand side, with a % prefix. Eg
+- For each CSV record, you'll assign values to one or more of the hledger fields to build up a transaction (journal entry).
+  Values can be static text, CSV field values from the current record, or a combination of these.
+
+- For simple cases, you can give a CSV field the same name as one of the hledger fields,
+  then its value will be automatically assigned to that hledger field.
+
+- CSV fields can only be read, not written to. They'll be on the right hand side, with a % prefix. Eg
   - testing a CSV field's value: `if %CSVFIELD ...`
   - interpolating its value:     `HLEDGERFIELD %CSVFIELD`
 
-- You'll be writing to hledger fields. (They can't be read.) They'll be on the left hand side, with no prefix. Eg
-  - setting the transaction's description: `description VALUE`
-
-- You can give a CSV field the same name as one of the hledger fields.
-  If you do, its value will be automatically assigned to that hledger field.
+- hledger fields can only be written to, not read. They'll be on the left hand side (or in a fields list), with no prefix. Eg
+  - setting the transaction's description to a value: `description VALUE`
+  - setting the transaction's description to the second CSV field's value  `fields date, description, ...`
 
 ## `fields` list