Dragon NaturallySpeaking

I saw an advertisement for this voice-recognition software.  I was intrigued and was thinking of asking Mike V., a relatively young whippersnapper who is en rapport with the latest gadgetry.  (When he visits my house he makes fun of my Jurassic electronics.)  But then I wondered how useful such a speech recognition application could be to someone who writes about arcane topics and uses high-falutin words.  Would the spoken 'animadversion' display as 'animal diversion'?  Would 'transcendental deduction of the categories' appears as 'transcontinental deportation of catnip'?  Would 'inverted qualia objection' show as 'involuted quails of Omaha'?

This morning I was pleased to hear from our old friend Vlastimil Vohanka who is also wondering about the utility to philosophers of DNS.  He conveys a remark by Baylor philosopher Jon Kvanvig on the latter's Facebook page:

"newest writing venture: I'm now composing using Dragon Naturally Speaking. It is amazingly accurate, beginning by trolling through everything on my hard drive to find appropriate vocabulary. So it recognizes 'Chisholm', 'Fregean', 'Chisholmian', as well as all the standard vocabulary in epistemology. And writing is so much faster: 5000 words in about 2 hours."

This is amazing. It implies that, were I to speak my neologism 'paleologism,' it would recognize and reproduce it and not spit out something like 'pal of loggers.' 

 

Vlastimil wanted my opinion of the DNS software, but I haven't used it so cannot comment.  If any reader has used it and wants to comment the ComBox is open.  Comments are moderated and won't appear until I have approved them.  Any that are off-topic will be summarily nixed.

Comments

[Francis Szarejko]

I've used DNS for several years in writing theological papers, taking notes on philosophical, theological, and apologetic works. It takes a bit of training for some of the unusual and technical language but after a few sessions I was dictating far faster than I can type with about 98% accuracy.

For me the only negative is that when I really need to think hard about something I am writing, I seem to think better with a pad and pen. But, then, I can dictate it in DNS when I need to.

Thanks for your work on this blog. Fran Szarejko

[Bill Vallicella]

Thanks for your comment and for reading the blog. All the best for the New Year.

[Doug]

I work for Nuance. There are two "issues" with DNS:
1. As Francis mentions, it won't work well unless your speech is more-or-less fluid. A lot of disfluency will just confuse it.
2. Its performance depends as much on your adaptation to it as on its adaptation to you. Both will take time. And there is no guarantee that after that time you will find its performance acceptable. (There are some people who just can't seem to find the sweet spot). But that disappointment is the exception to the rule.

It really does work well for most people for dictation: if you have a good idea what you are going to say, you can likely get it down much faster with DNS. You might even find editing/correcting faster with DNS, but you might prefer to do that the "old-fashioned" way.

[DrDoctorDr]

I am on sabbatical this semester, and with a number of writing projects to complete, I thought I would give this product a go to improve my productivity. Initial results are promising; I roughed out a draft of about two pages to a book review in about five minutes. The trick is to know what you are going to say ahead of time, and then just say it. A little odd speaking the punctuation; it really is like dictation.

The Home edition will only transcribe live dictation. I purchased the Professional version, which does live dictation, but can also process recorded dictation. I borrowed a digital voice recorder from our tech people, so I can "write" whenever I want by dictating to the recorder, and later have it transcribed by the software. The Professional version I bought was on sale for about $90 at a big-box office supply store, and included a headset with a built-in microphone. With that, the accuracy is extremely good.

It will take some adaptation of writing-style, but I can see myself getting a lot more text on the page by this method, where I can "write" even when I'm going for a walk.

[Vlastimil Vohánka]

Thanks for this, Bill.

Here's what Alex Pruss wrote. Jon Kvanvig's "5000 words in two hours translates to 42 words per minute. I can type at 80-100 words per minute, though in practice I write much more slowly because I have to think about what to say. And it seems much easier to write than to speak, at least for me. I am sometimes tempted not to talk to family members but to type to them."

If it's more easy for one to write (carefully) than to talk, typing still may well be faster than using DNS. Francis said in a comment to this thread that one can still dictate in DNS what he has already pondered on. Still, many times I had to refine my sentences about, say, a philosophical issue over and over. And I doubt I'd do better if not typing but composing only in my head instead. But it's hard to decide whether DNS would be helpful for me so long as I haven't tried the SW.

[Bill Vallicella]

Doug,

Thanks for your frank assessment.

Dr,

Your comment is extremely helpful. It has no problem with theological terminology?

Vlastimil,

Thanks for the 'Prussian' report!

[Kevin Wong]

I have used the built in voice recognition software on Windows 7 on the occasions that my wrists are in pain from my tendonitis. On a scale of 1 to 10, 10 being the best, I give it a 7. It gets a C since it has a bit of an interface problem. I would be interested in giving Dragon a try.