The golden rule is that the indent should be as wide as an imaginary perfect square described by the left margin, the baseline of the line above, and the x-height of the line below. The fourth side will then describe the indent.

Or one Em, which in most fonts should be roughly the same as the first method.

Either one em or one lead is standard. I prefer a little more, though, as do a lot of designers. Basically as long as you're under two ems you're safe.

One thing to consider is how the indent will play against orphans.
Another is line length.

I usually set the indent to equal the leading value, as this generally gives a good proportion that works for both narrow and wide column settings and adapts quite well even for very wide linespacing.

The other method is that suggested by Claus, creating an optical square, but this results in a deeper indent than I usually like, and the squareness ends up being variable depending whether the line below starts with ascender letters of all x-height.

thanx for all the reply's very helpful to know.

@Clauses, i didn't quite understood what you mean with "he indent should be as wide as an imaginary perfect square described by the left margin" English is not my native language so that probably has allot to do with me not understanding that sentence, could you maybe clarify a bit please?

Occasionally in old publications (early 20th century) and in typewritten letters from that period I've seen huge indents; like maybe 1/3 the width of the line. I was never sure if that was common in those days or just an oddity.

What Claus meant:

thanx for clarifying that Publishingmojo, appreciate the help people.

Carolina, as a standard rule, I set the indent measure as the same of the em dash of the typeface I'm using in the paragraph. I have exeperienced that, for example, in a Garalde set at 11 points, this measure turns out to be - approximately - at 3,5 mm or 4 mm. If it appears to be too big, I decrease it a little (from 4 to 3,5 mm.; the first line indent shouldn't be too big). This way, it usually works fine for me. I'm not totally sure, but I think I read this "em dash method" somewhere in a Jan Tschichold's essay.

I often use a deeper paragraph indentation in flush-left/ragged-right (‘unjustified’) alignment.

What is your thinking in this regard, Maxim? Without the anchor of consistent margin on the right, the big indent seems to make the first line float to the left, like a boat that has come unmoored.

The standard—one-em, or one-linespace—indent may be, and is, sufficient for a justified column. However, I believe that with the right-hand edge of the column ragged, a stronger paragraph break is necessary. The efficient and creative use of a deep indentation in Typographische Monatsblätter, and other print of the 1970s still looks very convincing to me.

with the right-hand edge of the column ragged, a stronger paragraph break is necessary.

Ah, makes sense now.

hhp

It doesn't make sense to me, yet.

What I'm wondering, Maxim, is why you think a ragged right edge makes a stronger paragraph break necessary? What is the relationship between the two? Personally, I'm having a hard time seeing anything in a ragged right margin that suggests the need for a larger than normal indent on the left, and the typography of Typographische Monatsblätter etc. strikes me as stylistic rather than functional in this regard (ironically, often the result of modernist experiments despite their designers' frequent appeals to functionalism).

[I'm also suspicious that the deep indent style had its origin in typewriter correspondence, and simply reflects the default tab setting.]

So what’s wrong with the typing format? A five-space-deep paragraph indentation does make a lot of sense in typing, with the right-hand column edge badly ragged because of the absence of hyphenation… So does double spacing after periods, btw.

Here are a couple of examples of deep indents in typewritten letters. Both are on White House letterhead — one from President Truman and one from President Roosevelt.

I guess Truman's was double-spaced because it was meant for his proof-reader? Come on, prez.

hhp

One more. Not as deep an indent, but it's more current — President Obama.

I'm still trying to understand what about the ragged right makes a deep indent on the left seem necessary to you. So far as I see it, there is no connection between the two, and the indent retains its primary importance in relationship to the left margin. The eye, reaching the end of the line, jumps to the beginning of the next line, which is to say to the left margin. Hence, the indent that is merely a visual indicator of a structural division of the text should be as deep as necessary to be unambiguous and no deeper. Unmooring the first line of a paragraph from the left margin with a deep indent means the reader has to go looking for it, instead of simply jumping to the left margin as usual. The overall visual affects of deep indents (and wide linespacing) on a page have dramatic visual impact, but they have nothing to do with the mechanics of reading, so don't serve the text or the reader.

The publishing convention for paper books is that the first line of a chapter, or a section after a line-break, is NOT indented.

Indents are to separate paragraphs, and since there is no paragraph before the first (by definition), no indent is needed.

I hope this is not to far off-topic, but all the examples shown have first paragraph indents. It may be a matter of preference; not indenting the first ¶ is mine. But whatever your preference, consistency matters.

@JamesM: Do you have a source for higher resolution versions of those letters? I’d love to incorporate them into a future presentation. If so, you can email me:

typographer@gmail.com

@Evie: See also:

Bringhurst, R. (2004). The Elements of Typographic Style. Point Roberts,WA: Hartley & Marks.

Tschichold, J. (1991). The form of the book. Essays on the morality of good design. (H. Hadler, Trans.). In R. Bringhurst (Eds.), Washington, United States: Hartley & Marks Publishers Inc. (Reprinted from Aufsätze über Fragen der Gestalt des Buches und der Typographie, by J. Tschichold, Ed., 1975, Basil, Switzerland: Birkhäuser Verlag)

> Do you have a source for higher resolution versions

Christopher, I don't remember the URLs but I found them via Google. Just do a Google images search for "typewritten letter president [name]" without the quotes and I'm sure you'll find them. Neither image was real high-res, but they were both bigger than the size I posted them.

(The reason I was searching for presidential letters is that I had a vague memory of seeing them previously with big indents.)

Update -- I found the URLs; see following post

I'm still trying to understand what about the ragged right makes a deep indent on the left seem necessary to you.

I wonder if making the indent larger than most or all of the "rags" sends the eye the message that "this is an intentional break, not an arbitrary byproduct of line lengths like those other spaces over there."

Found them, Christopher...
http://www.harmerschau.com/lotphotos_large/90/112306.jpg
http://myloc.gov/_assets/Exhibitions/naacp/worldwarii/Assets/na0095_01_e...

I wonder if making the indent larger than most or all of the “rags” sends the eye the message that “this is an intentional break, not an arbitrary byproduct of line lengths like those other spaces over there.”

Yes, something to that effect. It’s just that the uneven lines in FL/RR setting create a lot of visual noise on page… Especially loud when there are no word breaks, which causes a ‘hard rag’, or strongly uneven right edge of the column. Bordering on unbearable in those languages that have longer words than English. So a stronger signal may be necessary to indicate a paragraph break.

Okay, so I understand the reasoning now, but I don't find it compelling. Those 'other spaces over there' seems to me to have nothing to do with what is happening on the left margin, which is nothing out of the ordinary for normal reading. So why introduce something that is out of the ordinary? As for the ragged right producing visual noise on the page, I agree, and I think the deep indent on the left only increases that visual noise: it certainly seems to in the 1976 Olympics page spreads to which Maxim linked. I think one of the reasons why adding space between paragraphs and not indenting them became a popular solution for ragged right text is precisely because making a consistent left margin anchors the text better and reduces the noise. And, of course, Eric Gill's solution for ragged right setting in his essay on typography was to get rid of the paragraph breaks altogether and use a pilcrow instead. Again, I think this was to maintain a consistent left margin contra the variable rag.

Making the left margin more ragged than usual because the right margin is ragged seems exactly what one should not be doing.

> Indents are to separate paragraphs, and since there
> is no paragraph before the first (by definition), no
> indent is needed.

Another reason for not indenting the first paragraph is that visually it creates a weak start.

I rarely indent first paragraphs. Occasionally a client (or copy editor) will think it's an error, but after explaining my reasoning they've always dropped their objection.

> So why introduce something [large indent] that is out of the ordinary?

In some cases (not necessarily this one) it might be the designer thinking "this looks like the text in a zillion other printed pieces; how can I make it look a bit different?"

See:

Frase, L. T. & Schwartz, B. J. (1979). Typographical cues that facilitate comprehension. Journal of Educational Psychology, 71(2), 197-206. doi: 10.1037/0022-0663.71.2.197

Abstract
Written sentences often contain several meaningful components (e.g., causes and effects or events in a sequence). Preliminary studies of technical documents showed that typographically segmenting these components improved raters' judgments of the comprehensibility of the information. In the present paper, this segmentation notion is generalized, suggesting that phrase segmentation and indentation can be used to facilitate comprehension. Five experiments were conducted (with a total of 72 college students or technical aides) in which Ss verified sentences by reading complex information in several technical passages. Meaningfully segmented and indented text resulted in 14–28% faster response times than standard text. Both segmenting and indenting significantly influenced performance; however, once a text had been meaningfully segmented, the addition of indentation cues did not significantly affect response time. These data shed light on persisting issues in typographic design, namely, whether there is an optimal length for lines and whether justified margins are desirable. Such factors appear to be of minor cognitive relevance. The critical variable is whether the format results in a display of easily encoded units, regardless of length or neatness of margins. (16 ref) (PsycINFO Database Record (c) 2012 APA, all rights reserved).

So if I'm reading that correctly, response time improved if text was divided into paragraphs by either 1) adding space between paragraphs, OR 2) by using indents. But doing both (space + indents) didn't add additional benefit beyond just using one of them. And that makes sense to me. Am I reading that right?

…adding space between paragraphs and not indenting them became a popular solution for ragged right text…

Fashions in paragraphing have come and gone.
The practice John refers to was the initial default of html on the Web, for some reason in opposition to the then current style in print. That is how it acquired its present popularity, with some new-media cachet as well.

Here is a previous implementation of a similar style, from the innovative Ladies Home Journal early in the Progressive era.

@JamesM: That’s what the abstract is saying, but I am reading the paper now and it is significantly more complex. There are actually 5 different experiments, each with different materials and results. I will be writing a review of this paper, but there is a lot of data to interpret. It may take some time. I will keep you posted.

I think I read somewhere that web browser default style was based on CERN's tech. writing style.