Filler Words and Dialog

Filler words are words that writing blogs tell you to eliminate from your writing. Most of the time, it’s good advice. But not always.

Let’s start with some sample sentences.

If you need help, then ask for it.

If you need help, ask for it.

You know, I was going to write about interaction, but I’ve gotten totally bogged down in the way that “then” conveys emotion to me. Am I alone in finding that the first sentence contains exasperation while the second one is a simple statement? I can’t read the first one in my head without giving it a dose of impatient annoyance, while the second one sounds like a simple statement that could be delivered gently. And the only difference, of course, is the word, ‘then.’ Weird.

Ahem, but moving on, my point is that tighter writing is better writing and so most often, sentence #2 is the better sentence. The more filler words you can eliminate, the faster the read, and for today’s audiences, faster is better.

Except that interaction requires a reciprocal action or influence, a back-and-forth, and sometimes some of those filler words are good at making the relation between cause-and-effect clearer. Not always! But sometimes. Let’s take a look:

I tried to get away. He jumped in front of me. I turned and ran.

I tried to get away, then he jumped in front of me, so I turned and ran.

In the second story, the filler words show you the flow of the action, and how the actions influenced one another. In this case, (IMO) the more tightly-written story is not the better story. Some filler words are almost always bad (‘very’ comes to mind) but if the words you use enhance the sense of interaction in your story, of reciprocal movement, they may be worth keeping.

So, believable stories include interaction. They also include criteria #6, reproductions of speech. In other words, dialog. This is an important criteria, in my opinion, but also so obvious that I’ve got almost nothing to say about it.

A quick example:

We talked about what we wanted for dinner but couldn’t agree on anything.

I asked him, “What do you want for dinner?” but he said, “Oh, I don’t know, what do you want?” I answered, “Not again. Can’t you ever just choose?”

The latter is a more compelling story. To make your stories more believable, you should put words in people’s mouths. That said, in the example, I included conflict without even thinking about it, because conflict makes dialog interesting. A story where characters exchange several minutes of small talk might come across as true, but the truth can be boring.

Never let believable trump interesting. 🙂

Next time: Five criteria in one shot, because they all boil down to “weird.”

Advertisements

Interaction is Action

And days later… I finally show up again. Bad blogger, bad, bad, bad.

But I emailed my editor last weekend with a list of things that I intended to fix in my manuscript so that he wouldn’t waste time telling me to fix them and he told me he hadn’t started yet and gave the manuscript back to me to make the changes. Yay, yay, YAY! I made those changes and many, many more. Read the whole thing out loud to myself Monday through Wednesday—getting totally hoarse in the process—and finally sent it back him Thursday morning. I’m hoping he’ll still be able to get it back to me by June 1, but I’m definitely more satisfied with it, so that’s good.

So getting back to Criteria-Based Content Analysis– interaction, criteria #5, seems so obvious that I sort of skimmed over it in my original presentation about CBCA, But when I was thinking about writing this post, I decided maybe it deserved a closer look.

Two sample stories:

I came home and watched my favorite television show, Grimm, before going to bed.

I came home and chatted on the phone with my brother before going to bed.

Same number of details, same contextual embedding, so two equivalently true stories, right? Not if you’ve added criteria #5, interaction, to your toolkit. The second story should sound more truthful. For law enforcement, the subliminal reasoning might be that the second story contains a witness and witnesses make lies much easier to disprove. If you’re being deceitful, including other people in your story increases the risk that you’ll get caught.

But as writers, we don’t need to care about the subliminal reasons why interaction sells stories—the fact is, CBCA tells us that it does. A story where two people (or more) are interacting is more believable than one where a single person is on their own.

Doesn’t that seem weird, though? If I tell you, “I was home alone and I ate pizza for dinner,” why would that be less believable than saying, “my awesome kid was home for the weekend and we ate pizza for dinner”?

And yet—those two stories aren’t equivalent. One of them is a boring story, a “so what?” story. And the other is a Story. Okay, sure, part of that is the adjective, but even without the adjective, one of those stories makes you wish I’d stop talking (writing) and the other one makes you want to ask questions. (Well, maybe they don’t—I shouldn’t be speaking for you! But that’s how I’d react to those two stories!)

The difference between them is interaction.

Google defines interaction as “reciprocal action or influence.” I’m thinking of it as back-and-forth, cause-and-effect. Interaction is dynamic—movement by one causes movement by another. That movement can be concrete—“I pushed him so he punched me”—or it can be… well, less concrete. If your character says something and another character’s feelings get hurt, it might seem as if no action has happened. But it has. Action doesn’t have to mean explosions and drama–as long as you’ve got interaction, and reactions with weight, you’ve got action.

One of the books that I started this weekend was published by a small press. It was… well, some editor somewhere thought it was good enough to publish. And a list of people on the first page claimed to have edited, copy-edited, and proofread it. But it was unbearable reading. My stepmother gave it to me with, I think, a subliminal “you should try to find a publisher, look at what’s getting published!” message. The reason that it was unbearable, though, was because it lacked interaction, except in the most superficial way. For the entire first chapter, the heroine was alone. She had a brief conversation with a butler—which might perhaps have been her story editor saying, “you need a conversation here, something needs to actually HAPPEN in this chapter”—but the conversation didn’t work as interaction because it was irrelevant to the story. It wasn’t a reciprocal back-and-forth, cause-and-effect: he was just a stage prop, there to open the door.

And, OMG, it was tedious.

I sort of figured out the importance of interaction when writing A Gift of Thought. Dillon’s scenes were such a challenge. As a ghost who couldn’t really communicate, he spent all his time watching and thinking, and his scenes kept feeling flat and dull. I didn’t have an explanation for it at the time—I just concluded, “Don’t write ghosts who are alone all the time!” but now I know that what was missing was interaction.

One piece of advice given to beginning writers is to start as close to the action as you can. Newbie writers, including me, start in the car on the way to the important meeting. (A Gift of Ghosts, anyone?) Or waking up in the morning. Or at home making coffee the morning before the plane crash that changes the character’s life forever. We start before the action begins, when instead we should be dumping our reader into the action immediately. But finding out where your story begins can be far more difficult than it sounds to people who aren’t writers. We need to realize that what makes action is interaction: your stories are active when you focus on your point-of-view character’s interactions with other people.

But wait, you protest! (Or at least I did, when I was thinking about this.) Lots of great stories have characters who are alone. What about The Hunger Games? Katniss is off in the woods by herself. Except that she’s really not. The moments when she’s alone are brief: she’s fighting the others, meeting up with Rue, searching for Peeta, finding Peeta, getting caught by Thrash—she’s hardly ever alone. All right, what about Castaway? An entire movie about a guy alone on a desert island. Nope. He interacted with Wilson, the ball. That was almost the point of the movie. What kept him sane was interacting, even when it was just interacting with his imagination. My Side of the Mountain? Again, no. He’s interacting with the hawk, with the kid from town, with his visitors. The Old Man and the Sea! Honestly, I’d almost give this one to my imaginary opponent in this discussion, except a) he’s interacting with the fish and b) what an insanely boring book. It proves the point: interaction makes stories interesting. Lack of interaction makes readers enjoy pleasant snoozes.

Takeaway: if you want your writing to feel real and to interest readers, you will focus on interaction, avoiding ever leaving your characters alone and making sure that the interactions you write create cause-and-effect movement.

I’m not quite done with interaction yet. My next post (which will not be tomorrow, because R is graduating from high school tomorrow and might not be Sunday, because it’s his last day home before he goes back to Seattle, but will, I hope, be soon-ish) will take a look at how we express interaction and how filler words (so, then, etc) might not be as bad as conventional writer wisdom says they are.

 

 

 

Context and point-of-view

Before I move off the subject of contextual embedding, I want to write about one more way in which context is a useful tool for writers, and that is in how it relates to point-of-view.

Any aspiring author who spends time in critique groups or on critique sites gets point-of-view issues beaten into their brain. For many writers, picking a point-of-view and sticking with it is one of the fundamentals of good writing. Some best-selling writers violate those rules all the time (Nora Roberts), but we should probably wait to do the same until we’re regularly hitting the best-seller lists. 🙂 Still, the idea of maintaining a clear POV isn’t complicated and it’s not hard. Basically, if you’re writing in first-person or limited third, the only information you can reveal to your reader is the information that your character has available to them. Straightforward, right?

You’re seeing out of your character’s eyes, hearing with her ears, smelling with her nose… and perceiving the world from the context of her brain.

I hate describing settings. I’m not a visual person, I don’t have a good memory for sights, and I don’t tend to notice a lot about the space I’m in. If someone stole into my house and turned all my pictures upside-down, it would take me weeks to realize what had happened. (Well, I’d probably never realize, because it wouldn’t occur to me that someone would do such a strange thing. But it would take me weeks to see that my artwork didn’t look right.) And it used to be that every time I hit a place in a story where I thought, “Ugh, new setting, I have to describe it,” I got stuck.

Writing one paragraph of description would take me about the same time as three pages of dialog and sometimes much longer. Those spots for me were dead spots and most of them in every book have been edited and rewritten and edited and rewritten some more. One of my fundamental rules as a writer is to skip the boring stuff. If I would start to skim as a reader, then I don’t want to include it in my books. I try to follow Kurt Vonnegut’s advice–every sentence should either advance the plot or reveal character.

But describing the setting is our essential contextual embedding, right? For a reader to feel that a story is “real,” they have to feel grounded, they have to have some idea about where they are, what the place is like. True. But if you approach your setting from the context of your POV character, you can also use your descriptive sentences to reveal character.

Let’s look at an example from A Gift of Thought:

“He smiled down at her, putting his hand over hers as they walked toward the low steps that led into the building. Automatically, Sylvie assessed the space. Three, no, at least four stories, with what looked like an open balcony on the front of the fourth floor. Multiple doors in the front wall meant too many entrances to easily defend, while pillars every ten feet or so could be useful hiding places or annoying visibility issues. On the left, the sidewalk sloped and the portico became a patio, a dead end unless you were willing to jump the railing to the street below.”

That paragraph once read something more like, “The building was x, y, z.” It was a new setting. I needed the reader to have some sense of the place, specifically its size and exits for when Rachel disappears. But the words were flat and dull and boring. I could barely read them myself without starting to skim.

The solution was to approach my contextual embedding from the context of Sylvie’s brain. Instead of simply describing the building, I tried to look at it as Sylvie would see it, and Sylvie sees everything tactically. As a Marine and a bodyguard, she cares about line of sight and exits. Now, all of a sudden, my description is still conveying the essential information I needed to get across, plus it reveals Sylvie’s character (or reinforces it, anyway), plus it becomes–for me at least–a much more readable wall of information.

For experienced writers or people in MFA programs, this all might seem completely obvious. Of course you can only know what your point-of-view character knows. But seeing with your POV character’s eyes also means thinking with her brain, noticing with her background, observing from the context of her past. A guy on reddit put it really well once–and alas, I cannot find his name and will paraphrase him badly. But he said something like a rose bush with one rose on it can be either a delightful surprise, a last glimpse of summer, or a sad survivor of the ravages of autumn. The choice should give the reader insight into the POV character, not into the author.

Tomorrow: Interactions and how “true” in CBCA terms and “right” in writer terms sometimes collide. Well, or maybe Monday. The household chores are piling up and my allergies will get a lot happier when I get rid of some of the dust in my house. 🙂

 

 

Contextual Embedding

Contextual embedding, criteria #4, is my favorite. So what is it? Let’s take another look at Story #3. (You’re going to get sick of this story–I’ll be mentioning it a lot before we’re through.)

Back when I was living in Oakland, my house was burglarized. I was living with my brother, his wife, her sister, their two dogs, and my five-month old son and in the middle of the night, someone broke in and cleaned us out. The worst part for me was that he or she stole the camcorder that I’d been recording my son with. I’d actually caught my baby’s first laugh on tape, and the thief stole it. Other stuff, too, but it’s the laugh that hurts.

This story uses multiple elements of contextual embedding. I tell you where the house is–Oakland, California. I tell you about the people who were there–my brother, etc. And I give you enough information to provide a sense of a specific moment in time–when my son was five months old.

Contextual embedding sets a scene. It provides context for the story, by including information about the place and time where an event happened. With CBCA, contextual embedding has a 69% success rate in determining a true story from a false one. After lots of details, the criteria with a 98% success rate, contextual embedding is one of the strongest.

Of course, using contextual embedding in your writing might seem obvious. All stories need a setting, after all. But you can be blatant about your contextual embedding–for example, starting a chapter with a heading that gives a place and a time–or subtle. And using subtle contextual embedding will make your writing richer and more believable.

Let’s look at another example, this time from A Gift of Ghosts. (Henry and Rose are the speakers.)

“Tommy Shaw put a garter snake in Rose’s lunchbox one time. We must have been about thirteen, fourteen years old.”

“Thirteen.” Rose shuddered. “It was my brand-new Hopalong Cassidy lunchbox, and I was so proud of it. When I opened it up and that snake slithered out, I cried.”

There’s some obvious contextual embedding in that quote. Henry and Rose were thirteen or fourteen years old, which tells us something about the timing. But there’s also some subtle contextual embedding in the shape of that lunchbox.

Now, I could have described the lunchbox in many different ways. I could have used no adjectives at all, simply said, “When I opened my lunchbox up and that snake slithered out…” Or I could have stuck with only “brand-new.” Or I could have described it. “It was my brand-new lunchbox, bright red metal, and I was so proud of it…”

But making it a Hopalong Cassidy lunches turns it into a detail that provides contextual embedding. It sets the scene at a specific moment in time–a moment when people knew who Hopalong Cassidy was. (Specifically, the year was 1953 and the reason Rose has a Hopalong Cassidy lunchbox is that it was the only branded lunchbox available at that time. Random trivia: Hopalong Cassidy was the very first branded lunchbox.) But if the lunchbox had been a Scooby Doo lunchbox or a Star Wars lunchbox or a Spice Girls lunchbox, it would have served equally as well as contextual embedding because any of those would have provided information about the timing of Rose’s life.

Stephen King once said, “Making people believe the unbelievable is no trick; it’s work. … Belief and reader absorption come in the details: An overturned tricycle in the gutter of an abandoned neighborhood can stand for everything.”

A lunchbox that is “brand-new” or “bright red” is still a detail. But it’s not the kind of detail that stands for much. A Hopalong Cassidy lunchbox offers more. For those who don’t recognize it, it offers the subliminal, ‘that must have been a long time ago.’ For those who do recognize it, it tells the era, it suggests that Rose’s family had money since they had a television in 1953, it suggests that Rose was probably a little spoiled, since she was proud of her trendy lunchbox. As details go, it is a very hard-working detail–and that’s because it’s a detail that provides contextual embedding.One more example, this time excerpted from A Gift of Time:

Travis paused, looking down at the boy. “Anything happen while I was gone?”

The boy dropped his head. …

“Heard a big splash, that’s all. Got scairt.” The last word came out in a mumble.

“Told you before, you’re way too big for a gator. It’s more scared of you than you are of it.” Travis made no move to get into the boat.

“You saw that big one. Thirteen foot long, it was! I’d be, like, breakfast. And not a good breakfast, neither. Not bacon and eggs, I’d be like a bowl a’ cold cereal.”

That alligator and that splash? And also the bits of dialect in the boys’ voices? They’re contextual embedding, details that evoke a setting, hint at a place and a mood. That line could have been, “It’s just scary out here in the dark.” That would have been a detail, too–the dark–and it would have made his mood clear, but it wouldn’t be nearly as successful at providing contextual embedding and making the story richer.

So, use contextual embedding to choose the details that will make your writing stronger. Stephen King’s overturned tricycle carries as much meaning as it does because it’s located in a specific place–in the gutter, in an abandoned neighborhood. It’s contextually embedded.

 

A Tale of Three Stories

CBCA measures the truth of a story using 18 criteria, but for writing purposes, not all the criteria are equally useful. For example, #15 is lack of memory. People telling true stories are more likely to admit that they don’t remember everything, that parts of the experience are fuzzy. But that’s only useful in writing if you’re using a first-person narrator, which I never do. Some of the other criteria are specific to law enforcement, so I’m going to focus on the criteria that are most useful to writers.

I’ll start by telling you three stories.

Here’s the first:

I was burglarized once. The thief cleaned us out.

Here’s the second:

I was once living in a house that was burglarized. I was asleep at the time of the robbery. The thief cut a hole in a window, unlocked it, came in through the window, went through the house and took everything portable—cameras, laptop computers, cash, camcorders. Then he left through the front door. He must have had to make two trips, he took so much stuff.

Two stories. One of them, according to Criteria-Based Content Analysis, should sound more believable to you. Obviously, I can’t read your mind. I can’t know what you think. But if you were a police officer listening to my stories using CBCA, you’d conclude that #2 was more likely to be true than #1.

The more details a story has, the more likely it is to be true. Details, as a criteria, measured by quantity, has a 98% success record in the tests of the viability of CBCA as a law enforcement technique. 98%. The more details a witness provides, the more likely it is that their story is true.

But it doesn’t stop there. Time for story number three.

Back when I was living in Oakland, my house was burglarized. I was living with my brother, his wife, her sister, their two dogs, and my five-month old son and in the middle of the night, someone broke in and cleaned us out. The worst part for me was that he or she stole the camcorder that I’d been recording my son with. I’d actually caught his first laugh on tape, and the thief stole it. Other stuff, too, but it’s the laugh that hurts.

Story #2 and Story #3 have roughly the same number of details. If the only thing that mattered was the number of details, they would be equivalently believable. According to CBCA, they’re not.

Story #3 should seem more believable, because not all details are created equal. Factual details matter. The thief came in through the window, etc. But factual details don’t resonate. People don’t respond to them. And they don’t make people feel the truth of your story. So what does?

Tomorrow: contextual embedding

Criteria-Based Content Analysis Or How To Tell Good Lies

At some point along my writing journey, I realized that I didn’t care all that much about being a “good” writer. There are so many good writers whose work is tedious to me. I was an English major. Many good, even great, writers bore me silly. James Joyce, yes, I am looking at you. Melville, ditto. Charles Dickens–could you possibly take longer to get to the point? (I excuse him, though, he was paid by the word.)

I decided that what I wanted to be was a good storyteller. And I realized that good storytellers were basically just good liars. Lots of stuff goes into a good lie in person–body language, eye movements, voice stress. But in writing, what makes a good lie?

Criteria-based Content Analysis is a technique used by law enforcement worldwide to determine whether an interview subject is telling the truth. It was developed in the 1950’s by a psychologist named Udo Undeutsch who hypothesized that “an account derived from memory of a self-experienced event will differ in content and quality from an account based on fabrication or imagination.”

The technique takes a statement from an interview subject and analyses it, looking for specific markers (aka criteria) that indicate whether the statement is true or false. No single marker is sufficient to determine truth (although there’s one that’s better than all the rest) but the more of the markers or criteria that a story includes the more likely it is to be true.

You know how polygraphs aren’t acceptable evidence in a court of law, because they’re really not very accurate? Criteria-Based Content Analysis has been accepted as evidence in the German court system (which is where it was developed) since 1954. Worldwide, CBCA is reported to be the most widely used “veracity assessment instrument.” So what better way to learn how to tell a good lie than to look at the markers that indicate whether someone is telling the truth? Once we know what those markers are, we can incorporate them into our work and tada, more believable stories.

Tomorrow, the first marker.

The Ghosts of Belize

Back in January, before I got swept up into ALM, I was working on a story called “The Ghosts of Belize.” I’d spent several days working on an outline for it, mapping the whole thing out, trying a new technique of outlining, but it didn’t work for me. Today I opened that story up. I already have the cover, so thought I’d go back to it while I wait for ALM to get back from the editor.

It’s the first time I’ve let a story sit for a really long time without working on it but with the expectation of continuing it and it was a great experience. Re-reading it, I could see exactly where I was going wrong. I was trying to write it like a Nora Roberts book. For good reasons–she’s an entertaining author who sells a ton of books, and I’d decided to take writing more seriously and pay more attention to the market and all that.

But I can’t write that way. Or at least when I do, it feels stiff and unnatural to me. I had lots of description, lots of scenery, room descriptions, etc. For one of the first times, my sense of the setting was very clear–because I’d worked on it a lot to make it clear. But it didn’t interest me at all. Also, I was trying to make Akira’s experience of late-onset morning sickness very real. It was, because it’s a familiar experience to me–I spent the last couple of months of my pregnancy throwing up ten times a day. But story-wise? Boring. Boring. More boring. She’ll still start out nauseous because it’s funny there, but she’s going to start feeling better almost immediately because it’s not funny or fun on an ongoing basis.

I read a blog post today, 5 Self-Publishing Lessons I Learned Between Books #2 & #3, by Molly Green, who says that her first lesson was that she figured out what she writes. I really need to do that. Realistic settings and realistic experiences are all well and good, but I don’t think they’re me. Real characters in unreal, entertaining situations maybe.

I’m going to continue working on Ghosts of Belize for a few days and see how it goes. It may wind up being a true short story, under 10K words and that will be fine, as long as it manages to be entertaining. That’s where it was going wrong in January. It was dark and not fun and that’s not me.