Novel diagnosis–voice revisited


Novel Diagnosis Series


Voice chart

This is a chart showing the variation in length of sentences. Green is for Lizzie Bright, page 33; pink is for Westing Game, page 33; and orange is for one of our readers, PB, page 33. (Waving to PB! Thanks for listing your sentence lengths.)

Last week, I suggested that a possible diagnostic for voice might be to look at sentence lengths. A couple people thought I was suggesting just looking at average sentence length; instead, I’m more interested in the variety and from this chart you can see the variety.

Lizzie Bright (green) has an average sentence length of 10.7, with a variation from average of -8.7 to + 16.3.

Westing Game (pink) has an average sentence length of 12.0 with a variation of -10 to +27.

PB (orange) has an average sentence length of 8.85 with a variation of -6.85 to +8.15.

So What?

So what, indeed? Does this simple test give us an idea about voice? Yes and no.

Certainly Westing Game and Lizzie Bright have a wider variation than PB. But to use an art analogy, there are different palates. One artist might choose a pastel palette and do a stunning work, while another chooses a bold palette for their work. I think PB’s voice palette might be more toward the pastel side and only she can decide if that’s appropriate or not.

Certainly, if the sentences vary from the average by only 1-2 throughout, then the voice is probably flat! I’ve read those eighth grade essays! But even taking on that challenge, a skilled writer might make that work.

Still Searching

One thing I try to do in my Novel Revision retreat and other teaching is to give the writer a tool that will help them see what needs to be done next. I’m not convinced that this tool is the most useful for voice. Voice is more about how the intonation plays out across a passage, and that involves so much more than just sentence length. I had hoped that this would be a simple test that would point out larger problems, but there are too many holes in it. So, I’m still searching for ways to make voice visible, easier to see the weak spots and easier to work on directly. Just haven’t found it yet.

  • Pam Beres
    July 23, 2007

    LOL–love the art analogy, and I agree–the voice in my novel would certainly be considered more “pastel” than “bold.” Hmmm, something to think about though.


  • darcy
    July 24, 2007


    Yes, something to think about. Of course, bold vocabulary could tilt the “palette.” We’re only looking at sentence lengths and there’s lots of ways to adjust the palette.


  • Kris H.
    February 20, 2011

    Wow, did I have fun with this. Although I didn’t graph it, I used Xcel to find my range, mean, median and mode (very quick with the “sort data” function). I two different analyses, one on running a page with mostly narrative and one on a page with mostly dialogue. For me, dialogue had the greatest range–from 2 to 25. Narrative ranged from 5 to 31. I’m hoping these ranges mean variation and not schizophrenia!