# How do you pronounce letters generated with \Bbb and \mathbb?

I'm writing LaTeX to speech software.

In LaTeX, I pronounce R as "Boldface Ar".

How should I pronounce $$\Bbb R$$?

• "The real numbers" usually. (but in an automated context if you have no clue to meaning, "double struck R" or "blackboard bold R" would be OK – David Carlisle Jan 23 '17 at 10:36
• If the purpose is to help read and write LaTeX code, then it would make more sense to actually read out the markup. If the software is intended as a format (like pdflatex or xelatex), then you might want to consider accsupp (or glossaries-accsupp) so the author can provide the appropriate pronunciation. – Nicola Talbot Jan 23 '17 at 12:30
• @NicolaTalbot From the research papers I have read, it is much more difficult for visually impaired people to interpret latex through code alone. – Anon Jan 23 '17 at 12:35
• I thought that might be the case. I'll add an answer that uses accsupp as I think that's the best approach to this problem. – Nicola Talbot Jan 23 '17 at 14:01
• @NicolaTalbot You should have told me 6 months ago when I started the project :P I'll have to see what it renders out. Mine is fairly unique in its approach, using sound effects and different voices to help articulate the formula. – Anon Jan 23 '17 at 14:04

BBB stands for blackboard bold, but "double struck" is also used (e.g. ℝ). However how it should be pronounced is another matter -- between reading out the markup ("dollar dollar backlash B B B space R") and the meaning of the symbol ("set of real numbers" is a big gap.

• This answer implies it, but to state it explicitly: presumably there are domains and fields in which ℝ doesn’t mean “set of real numbers,” which is going to be a big problem for this project. – KRyan Jan 23 '17 at 16:28
• @KRyan - What do you base your presumption on? – Mico Jan 23 '17 at 16:46
• @Mico Just because there is just about no symbol in the world not redefined by someone for some purpose. I can’t find any other usages of ℝ itself in a cursory search, but the point remains even if it applies to other characters. – KRyan Jan 23 '17 at 16:50
• @ChrisH Right, hence the large gap between “dollar dollar backslash B B B space R” and “set of real numbers”—choosing where to fall on that continuum for each symbol is going to be tricky. “Blackboard bold R” or “Double-struck R” are probably better choices than “the set of real numbers” since even though the latter is more useful in almost-all situations, in that one weird corner-case it will be wrong. – KRyan Jan 23 '17 at 17:24
• ...or, closer to the OPs situation, \mathbb{P}, which I have seen being used both as generic notation for probability measure functions and the set of primes. – phipsgabler Jan 23 '17 at 19:39

When I first read your question, I wasn't quite sure which of the following applied to your tool:

1. The tool is for visually impaired users writing LaTeX code.
2. The tool is for visually impaired users who have been sent a LaTeX document because the PDF generated by pdflatex doesn't read correctly when used with a PDF to Speech tool.
3. The tool is for document authors who want to create an accessible version of their document for visually impaired users in much the same way as document authors might want to create both a PDF and HTML version of their document for the convenience of their readers.

I'm guessing from your comments that it's more likely case 2 or 3.

In the first case, the user really needs to know the markup. Trying to interpret the LaTeX code in this instance is like trying to create an audio version of lyx that hides the commands. I'm not sure how well that would work for visually impaired users.

In the second case, there's no reason to suppose that the user who has been sent the LaTeX code actually knows much about LaTeX, so reading out the commands is more likely to confuse them. In this instance, I think the symbols should simply be pronounced by their name rather than what they represent. Others have already mentioned Unicode. Given its ubiquity, I think it's best to use the Unicode description. This helps with standardisation.

If the listener understands the subject matter, they'll know what the symbol represents, especially if the topic is at an advanced level.

Now suppose the document is an introductory text about set theory that has something like:

The set of real numbers (denoted ℝ) …

This won't make any sense if it's read as

The set of real numbers (denoted the set of real numbers) …

The listener needs to know what that symbol is in case they encounter it in another document with another text to speech system that doesn't try interpreting it.

It's far less confusing to hear the symbol pronounced by its Unicode description that it is to hear an incorrect phrase caused by a misinterpretation of the symbol by an automated system that can't recognise the context.

For example, suppose the tool tries to be clever and reads \$1 out as "one dollar", which is the way$1 is usually read by a sighted reader if it's in a currency context, but perhaps the document is actually an introductory guide to regular expressions and it's referring to the first matched group. The tool has now misinformed the listener.

The third case is the best use of your tool. The document author is going out of their way to provide an accessible version. If I'm writing a document that needs to be converted to multiple formats, I expect to write additional code that I wouldn't need if I was simply creating a PDF (except perhaps for trivial documents). As long as your tool recognises the accsupp package, the document author can provide the most appropriate pronunciation.

For example:

\documentclass{report}

\usepackage{amssymb}
\usepackage{accsupp}

\begin{document}

$\BeginAccSupp{ActualText={set of real numbers}} \mathbb{R} \EndAccSupp{}$

\end{document}


This provides the text to read out when encountering that symbol. If you want to use the symbol a lot, then you might find it easier to use glossaries-accsupp:

\documentclass{report}

\usepackage{amssymb}
\usepackage{glossaries-accsupp}

\newglossaryentry{R}{name={$\mathbb{R}$},
text={\mathbb{R}},
description={},
access={set of real numbers}}

\begin{document}

$\gls{R}$

\end{document}


This internally uses \BeginAccSupp but it's less cumbersome.

Alternatively:

\documentclass{report}

\usepackage{amssymb}
\usepackage{glossaries-accsupp}

\newglossaryentry{R}{name={$\mathbb{R}$},
text={\mathbb{R}},
description={},
access={set of real numbers},
firstaccess={set of real numbers (double-struck capital R)}
}

\begin{document}
$\gls{R}$

$\gls{R}$

\end{document}

• This is both obvious and irrelevant. Of course document authors should provide accessibility support. That does a blind person absolutely no good when presented with a document that doesn’t, which is the point of the project and this question. – KRyan Jan 23 '17 at 16:29
• @KRyan I added this answer following on from the comments to the question since it's related and there's not enough room in the comments to elaborate. – Nicola Talbot Jan 23 '17 at 16:41
• Well, all right; that’s kind of an abuse of the SE format, and coming here from HNQs I was confused by it. As someone not from here, I would comment that I think that’s a bad idea and it would have been better to produce a self-answered Q&A to link to from comments. This information is useful, just not here, where it will be hard to find for the next person. – KRyan Jan 23 '17 at 16:46
• @KRyan - What does HNQ mean? – Mico Jan 23 '17 at 16:48
• @KRyan ... if you're reading on a big screen. There are plenty of mobile users who don't get the sidebar – Chris H Jan 23 '17 at 17:00

When i was a math student all these double struck letters were called "Japanese", because they were strange and "nobody in this class is knowing japanese 'letters'". Moreover this is an unambigious naming.