A quoted message block

Files and text

A run of earlier text reproduced inside a reply and marked line by line so a reader can tell the old writing from the new.

The shape

Quoting is a typographic convention, not a mechanism. Its whole job is to make two voices legible in one block of plain text where no styling is available. The convention settled on a marker at the start of every quoted line, because a marker at the start survives rewrapping and stays visible however narrow the column becomes.

Marker

a greater than sign at the start of each quoted line, conventionally followed by one space

Attribution

a single line immediately above the block naming the writer being quoted, usually ending in a colon

Nesting

one additional marker per level, so the oldest text carries the most markers

Blank line

one empty line, or an empty marked line, separating the quoted block from the reply beneath

Marked blank line

a line consisting of markers alone, preserving a paragraph break inside the quote

Line width

set by whoever quoted, since quoted text is normally rewrapped as it is inserted

Ends with

nothing. There is no closing marker, so the block ends where the markers stop

An earlier writer wrote:
> the first line of the text being quoted
> the second line of the text being quoted
>
> a new paragraph inside the same quoted block

the reply starts on this line

An earlier writer wrote:
>> the oldest text, quoted twice over
>
> a reply to it, now itself being quoted

the newest reply

Where you meet it

Quoted blocks turn up in three places that behave differently. In a message thread the client inserts the markers for you when you reply, which is why the shape is so consistent within any one thread and so varied between threads. In a pasted transcript somebody has assembled the block by hand, and the markers were typed like any other character. In an image of a conversation there are no markers at all, only pixels arranged to resemble them, and the shape rules here do not reach that far.

The middle case is the one worth slowing down for, because a pasted transcript looks exactly like a client generated one. Both are plain text. Both carry markers. There is no field, character or spacing difference that separates a block a client wrote from a block a person typed out.

Quoting also appears at a remove, inside text that is itself being quoted. A transcript pasted into a reply acquires a layer of markers on top of the layer it already had, so depth accumulates from two unrelated sources and stops corresponding to the age of anything. Where you meet a block four or five levels deep, some of that depth is usually history and some of it is packaging, and the two cannot be separated from the marks alone.

How to check it by eye

What a wrong shape means

Markers vanish easily. A field that strips leading characters, a paste into a target that reformats, a client that converts marked text into styled text and back again: any of these can remove markers from some lines and leave them on others. The result reads as if the quoter wrote part of what they quoted, which is a formatting accident and looks like a claim.

The opposite fault is a block where every line carries a marker including the reply, usually because the whole message was quoted as a unit somewhere upstream. Depth then means nothing, since it has been applied uniformly rather than accumulated.

Uneven depth inside one paragraph, such as two markers on one line and three on the next, generally means a client rewrapped the text after quoting it. The lines were split at new places and the marker prefix was reapplied inconsistently. The related damage, where line structure disappears altogether, is set out on the lost line breaks card.

A block with no attribution line above it is not malformed. Attribution is a convention that many clients omit, and its absence is a fact about the client rather than about the text.

What a correct shape does not prove

This is the shortest reasoning on the site and the one most often skipped. Every character in a quoted block was typed or pasted by the person composing the reply. The markers, the attribution line, the depth, the spacing and the quoted words themselves all arrive from the same hand. A quoted block is a person telling you what somebody else said, formatted so it is easy to read.

Formatting is therefore about legibility. It was never designed to carry provenance, it has no field for provenance, and a well formed block is well formed because a client applied a rule or a person copied a habit.

• The attribution line is ordinary text. Any name can be typed above any block.
• The nesting depth is a count of characters somebody added, and it can be increased or reduced freely.
• Quoted words can be shortened, reordered or altered without leaving any mark in the shape.
• Omission is invisible. A block that removed a sentence looks identical to one that did not.
• A perfectly consistent block may have been assembled by hand, since consistency is easy to type.
• An inconsistent block may be entirely faithful, since clients damage markers routinely.
• The timing of the quoted text is absent. Quoting carries no date of its own.

If you want text tied to a key rather than to a formatting convention, that is a different shape with a different envelope, described on the clear signed message card, and even there the card is careful about the size of the claim. Quoting has no equivalent, and adding markers to signed text usually breaks it.

What this card is not

Nothing here covers how to establish who wrote anything, because a reference of shapes cannot. The nearest thing on this site to a shape that links text to a holder of a key is the clear signed message card. Damage caused by fields that join lines, which affects quoted blocks along with everything else, is on the lost line breaks card.