Why can I only use some capital greek letters inside my equations?

For example:

$$\delta % ok \Delta % ok \epsilon % ok \Epsilon % error$$
• I also ran into this problem today. The capital letter $E$ is not what I wanted. I wanted the symbol $\mathcal{E}$.
– OSE
Commented Nov 11, 2013 at 0:40
• Don Knuth spent some time to explain the TeX logo. I think it still makes a good read. Ironically, what we call "Greek" letters should be called Byzantine script, as ancient Greeks (i.e., Archimedes and his likes never used it). However, the question is legitimate and dismissing it as "just write latin E" is a bad way to answer it. This does not explain, for example, what happens when you typeset a Greek book in TeX. Commented Jul 30, 2016 at 8:17

Uppercase epsilon is E.

$$\delta % ok \Delta % ok \epsilon % ok E % ok$$
• wow and i thought c++ was inconsistent
Commented Aug 9, 2010 at 0:34
• Yeah, as much as I love LaTeX for its typesetting skill, it makes so little sense internally. Commented Aug 9, 2010 at 0:54
• unicode-math has \Alpha, etc. defined, so this will be more consistent in the future. Commented Aug 9, 2010 at 14:54
• The uppercase epsilon is E, so I don't see how this is inconsistent. This visual similarity (identicalness, actually) is a feature of the scripts themselves: what would you expect from an "uppercase epsilon" if not the letter E? Commented Aug 10, 2010 at 22:08
• I think this is one to learn from for LaTeX3: on semantic grounds there should be a macro here, even though in the default fonts it doesn't show up visually. Commented Aug 12, 2010 at 19:28

To quote from The LaTeX Companion (p. 527),

Those capital Greek letters not present in this table are the letters that have the same appearance as some Latin letter. Similarly, the list of lowercase Greek letters contains no omicron because it would be identical in appearance to the Latin o. Thus, in practice, the Greek letters that have Latin look-alikes are not used in mathematical formulas.

So basically, nobody would use a capital epsilon in a formula because it'd be visually indistinguishable from E.

If this bothers you, you could define the missing macros for Greek letters, e.g.

\newcommand{\Alpha}{A}
\newcommand{\Beta}{B}
\newcommand{\Epsilon}{E}

and so on.

• I remember in a class on spinor methods discovering with some alarm that the basis spinors were denoted omicron and iota, with the two components denoted by super- and sub-script 0 and 1. Since the spinors were indexed by capitals, it made a page of $o_0^{A'} \iota_A^1$, handwritten, rather challenging to re-read. When the lecturer remarked in an aside that this was a rather exquisite in-joke, he only narrowly escaped being collectively throttled by the class. Commented Aug 9, 2010 at 8:37
• My measure theory lecturer wanted to use $\mathbb{C}$ for something other than the complex numbers, but we shouted him down. There were still at least 4 near-indistinguishable "M"s... Commented Apr 2, 2011 at 15:13
• Use \providecommand instead, that way it's forward compatible in case these commands are defined in the future. Commented Oct 1, 2013 at 5:13
• Except, of course, for the Beta function, the H theorem, and certain Gegenbauer functions.
– E.P.
Commented Jan 17, 2014 at 19:17
• I can't see the difference between $\upsilon$ and $v$. Commented Jul 30, 2016 at 8:28

David's point is worth underlining:

So basically, nobody would use a capital epsilon in a formula because it'd be visually indistinguishable from E.

If you wanted to type something actually in Greek, then you would use one of the packages designed for that (search on CTAN for "greek" to get an idea of what's available). The Greek letters that are defined in unadorned LaTeX should not be viewed as letters but as mathematical symbols. So α should be viewed in the same regard as something like ≤. This is underlined by their appearance: just as "x" looks different in text and in maths, so "α" will look different in text and in maths. So the fact that "Α" (\Alpha) and the like are missing is simply because regarded as a symbol, it's just "A", and no-one would ever write \Alpha because they can just write A and the fact of calling it "Alpha" adds nothing. Compare this with "x" (x) and "×" (\times) where the different names actually mean something (variable versus operator).

So, in summary, if you want to type \Epsilon and the system complains then you are doing something wrong: either you are trying to type something in Greek without loading a proper Greek alphabet, or you are distinguishing in the source code something that will be indistinguishable in the actual document.

Other answers have made this fairly clear: a capital epsilon is identical to a capital E, so there's no need for a separate LaTeX command. However, there is a point missing from the answers above. Roman letters (by which I mean letters from the Roman alphabet) are conventionally set italic in math mode, to make clearer the distinction between maths and text. But Greek letters are set upright by default (though some journals deviate), because there's no distinction that needs making (assuming you're not writing in Greek). So if you really want a capital epsilon, and you want a consistent look for all your capital Greek letters, you should use \mathrm{E}.

• That's not true. The Greek letters are typeset italic. There's a package to typeset them upright: upgreek
– TH.
Commented Nov 24, 2010 at 1:00
• Hmm, that's strange; mine are definitely upright, and I can't imagine that I'm calling upgreek without knowing it. Could this be a result of font selection? Commented Nov 24, 2010 at 22:20
• Check this at section 5. Commented Apr 12, 2011 at 8:47

Well, my wild guess is that such macros are missing because CM has no such glyphs to begin with, Knuth's Greek is meant for use in math and not for running text and thus there were no point in duplicating glyphs that are visually indistinguishable from Latin counterparts.