Writing Isn't Just Writing, Its a Lot More

This blog will come to an end soon. So, if the remainder of it sounds like a collection of odd topics rather than a continuing theme, that is probably because it is. I have been on a journey (I hate that phrase) for the last three years to improve my writing skills and become more familiar with independent publishing. I think I have made good progress. Although there is still much to learn, I feel that I am turning the corner from discovery to practice which will make you as nearly perfect as you are capable of being. I will pick up this thread later. However, for now I wanted to revisit a topic I first posted about very early in the life of this blog and delve more deeply into the assertion that there is a lot more to writing than just writing. I referred to the cliche fantasy that many people (writers and non-writers as well) have in which somebody types the first line "It was a dark and stormy night" and then continues, without any hindrances, to write the great American novel.

Anyone who can type or form legible letters with a pen and knows enough to put those letters together to form words can be said to be writing.  But does that make them a writer? Probably not. This would be equivalent to someone pecking out chopsticks on a piano and then attempting to say they are a musician.

What if our friend with the pen or typewriter can form sentences from those words which are intelligible to another person. Are they a writer now? Again, probably not. That would be like our friend on the piano following up chopsticks with Twinkle, Twinkle Little Star and trying to claim musician status at that point.

And we could go on. Being a musician requires precision, timing, nuance, interpretation, breadth of musical instruments, and a broad background in music. The Beatles, who are generally recognized as excellent musicians could not read music. But George Martin could. And one can reasonably wonder if the Beatles would have reached the levels they reached without George Martin. One can reasonably wonder the same about Brian Epstein and his skills in business and promotion. The point is that being a successful musician requires a whole range of skills sometimes found in a single individual, sometimes in a group. To be fair, we have to admit that nobody excels at all the skills needed to be a successful musician. But you do have to excel at enough of them to stick out from the crowd. Or you have to have a team made up of the right people with the right skills.

Being a successful writer also requires one to excel in a number of areas. You need to be familiar with what other writers have said and done. You need to be able to create compelling phrases and design compelling stories or arguments. You need to find your voice and your audience. Sometimes writers rely on editors, business managers or friends who read chapters or comment on ideas. Sometimes, they will turn what they have written over to a team at a publishing house who will compensate for what they do not know.

There is a tired old joke about a guy who calls a mechanic because his car won't start. The mechanic looks at the engine, tinkers with a few things and then pulls a hammer out of his tool kit which he used to smack something in the engine. The car owner tries to start the car and the engine roars to life. Then the mechanic hands the car owner a bill for $100. The owner is shocked by the bill and protests that all the mechanic did was hit the engine with a hammer. The mechanic points out, in his defense, that the $100 was not for the hammer strike but was for knowing where to strike. Yes, this is a really stupid old joke but it highlights the fact that when somebody is successful at something it may lead onlookers to believe that anyone could have done it, when, in reality, it was the product of a lot of background work and practice some of which may not appear related to the successful act at all.

When somebody shows up at a bookstore for a book signing, that book is the product of a lot of activities that may not appear to have anything to do with writing.  These activities include, but are not limited to reading, studying, writing practice, thinking, experiences, arguments with friends and all manner of efforts to develop your creativity and imagination. If you are an independent writer, then everything falls into your lap including formatting, graphics, and marketing. If you have ever had the experience of coaxing a wound-up toddler into taking a nap, you have a tiny glimpse of what this process is like.

I attended a workshop, years ago, in which the instructor was asked about writer's block and what to do about it. Even though it was actually decades ago I still remember her response. "If you have writers block," she pointed out "then you don't have anything to write and you need to go back to your preparations such as reading, thinking, being imaginative, writing something shorter, and so on." If you have writer's block, then you don't have anything to write. Brilliant! There is a huge amount of preparation necessary to bring you to the point where words start to flow. And failing to do the work leads to writer's block. Sometimes the preparation includes things like not thinking about what you are trying to write or engaging in activities that you don't normally do to jog your thinking.

We can think of writing as having three stages. One is all the preparation that you have to do to get the words to start flowing. The next is the stage where you let the words flow and try to get them on paper. And the third is all the work necessary to get what you have written into the hands of people who want to read it. I am reluctant to put numbers on this so I would say that the vast majority of the time goes into stages one and two. This means that the bulk of the time when somebody is doing the work needed to be a writer that will not be actually writing at all. 

Going back to the common fantasy, it is possible that someone was born with a gift for writing which was then developed in school. Then they had a collection of life experiences which coalesced into a story that other people would like to read or hear. So, after procrastinating for an acceptable period of time, they sit down at their computer and start writing. The word flow and the experience is unbounded ecstasy. After combing over a few drafts, they send it to an agent who loves it. The agent sends it to a publisher who agrees to publish it and it becomes a best seller. 

Does this ever happen? It probably does but it is extremely unlikely. People become millionaires by buying sweepstakes tickets. And making your fortune by winning the sweepstakes is not what I would call a solid career plan. You might get lucky once but the odds of getting lucky twice are (extremely unlikely) squared. So, if you are planning on being a writer and spending your life writing, you should probably come up with a better plan.

Can AI Write Stories ?

A growing concern these days is how much, that we currently do to earn our livings, can potentially be done by computers. Many traditional professions have already been impacted by the encroachment of computer technology. In fact, professions such as accounting, medicine, law, and education have already been impacted to the extent that they are no longer recognizable from 50 years ago.  Much of the drudge work has been automated so that the number of jobs is a fraction of what it used to be. Over the next 50 years it will become even more unrecognizable and require fewer and fewer people. Warehouses and restaurants are using more automation thus requiring less people. Intelligent vehicles will put a big dent in the number of jobs where people drive for a living. And so on.

So, for people, who write for a living, the question arises - generally, can computers replace writers, and more specifically can AI write stories? The answer, as is the case in most threatened professions, is yes and no. Before you get threatened or elated, let me explain. I'll cut to the chase, so you can see where I am going. If the story is based on a template, the answer is yes. If the story is based on reporting and sharing a person's internal subjective experiences, the answer is no.

I'll take the No side first and then go back and forth a bit. Computers do not have internal subjective experiences. That is to say that computers don't have emotions or feelings. Although I cannot say that computers will never feel, there is nothing on the horizon that suggests, in even the smallest way, that they will ever be able to experience feelings. They may be able to act as though they are responding to feelings. But they won't actually experience them. I have to say that it is a mystery to me how people feel. But they seem to. So, we just accept it. Since machines are not capable of feelings, and stories report on feelings, then anything not capable of feeling must not be capable of reporting on them. Hence, they cannot write stories.

So, let's take the other side. If a story is based on a template, then it may be possible to have an intelligent computer fill in the template and create a story. This would probably start out very crude. But if we had software that creates stories, and then have those stories evaluated by people, we could use machine learning to determine what makes the stories good and, thus, improve them. Theoretically, over time, the computer would get better and better at writing stories. 

I would point out that when information systems were first introduced into corporate environments, the firewall between human and computers was judgement. People are capable of making judgements, the reasoning went, computers are not. Of course, today, computers make judgements every day at a rate which is many orders of magnitude greater than the rate at which humans make judgement. In fact, every time you react positively to an ad, email or popup while browsing the web, a computer somewhere has successfully predicted your judgement. 

So, this is not as farfetched as it seems. There were studies where music was generated by computers and then evaluated by people. People were suspicious of this suggesting that no computer could write music as good as Beethoven (which tends to be more emotional than, say Bach which is more technical). But empirical studies suggest that this is not true. So, if computers can create music as good as the music masters, why couldn't they write stories as good as the literary masters? If you are still doubtful, try asking yourself if this could happen in 10 years, 100 years, 1000 years and so on. I think you will find that the answer is not if computers can do this, but when. And when often turns out to come sooner than more people would believe.

If there is an answer to that question (can AI write stories) I think it would be in the fact that the way people feel about the world changes over time and any author writing a story would be writing to the audience of readers at that time. Computer algorithms would have to be constantly updated. Since the algorithms would get better with refinement, the changing target of current emotions might present a challenge. Granted many stories written for an early 19th century audience still appeal to readers in the 21st century. But many of them don't. Even for those that do, they appeal to a more limited audience. Ask anyone if they would rather read Jane Austen's Emma or see the movie Clueless. I think the votes for Clueless would dwarf the votes for Emma. Similarly, ask someone if they would rather read Dickens Bleak House or watch the BBC series and I think the results would be the same.

This leads to a hopefully not so unexpected conclusion. Perhaps computers will be able to produce popular stories such as the script for a Situation Comedy or a Blockbuster movie, while deeper more reflective works will be left to humans. This is right in line with the professions that I began this piece with. We will need vastly fewer writers and those we have will be very skilled and very talented, which is to say that computers will take jobs away from average people. Hopefully, new average jobs will be created to replace them.

 

Save the Cat as a Design Template

Following our analogy between writing software and writing stories, we can gain additional insights by exploring the use of design templates. 

Back in the early days of information systems development, say the early1970's, nearly all software was written a line at a time using a programming language like COBOL, P/L1 or C.  Over time developers realized that certain functions such as a sort routine or a mathematical function were being written over and over again, by each programmer who needed them, leading to a great variation in software reliability and added expense due to the enormous duplication of effort. Eventually, these common functions were collected in software libraries. Fast forward to today and one of the most popular programming languages, called Python, allows software developers to do a small amount of traditional programming while pulling the vast majority of needed functionality from existing libraries. I hope I haven't gone into too much technical detail here. But, I wanted to show how you can go from writing everything from scratch to assembling it from existing reusable components. And I wanted to show this because the same thing is possible for writing stories.

There are simple templates for writing stories that provide heuristics for the major components of the story such as: You want the reader to be able to identify with the protagonist. All primary characters must be believable people. You must keep the reader interested by revealing the plot in a way as to keep their attention. And so on. And while those are helpful, they leave a lot for the writer to do. This is not a bad thing, as the writer should be contributing a lot to the story. But, there are stories written for different purposes. I mean no disrespect by saying that there are junk food stories to satisfy some readers who want entertainment and emotional stimulation. And there are haute cuisine stories for people who are looking for something deeper and more satisfying. For junk food stories, it is way too much effort to write them from scratch, and more sophisticated templates are needed to leverage the writing process.

There is already at least one reusable template for a story which is widely used. It is called Save the Cat by Blake Snyder, and it was introduced for screenwriting. It focuses on movie and TV scripts but has been applied to novels and other story vehicles as well. Having a template like Save the Cat provides several benefits. It provides a structure for writers so they don't have to develop their ideas from scratch. It provides a template to facilitate the evaluation of scripts so the people who buy scripts can go through them more quickly. And, it provides the audience with what they are expecting.  

The Save the Cat template is extremely popular at present and its advocates make it sound as though anyone who does not follow it will never get any scripts accepted or novels published. This may very well be true for the popular market but there are two additional things that one must consider. First, not all writing targets the popular market. Some writers wish to achieve something different, perhaps more literary. Perhaps they want to try something more novel, no pun intended. Maybe they want to tackle one of Jane Austen's novels in the style of Henry James. Maybe they just want to write what they want to write, popular markets be damned. 

The second problem with the popular market is that it is fickle. If you followed television from the middle of the 20th century you would see detective stories, followed by cowboy stories, followed by stories about hip socially aware characters. When detective stories rule, just try to get a cowboy script published. As a thought experiment imagine trying to get a popular show from the 1950's accepted in the 1970's, or vice versa. Would the Beverly Hillbillies sell today? Would All in the Family have attracted an audience in the 1950's? 

The problem with chasing popular tastes is that popular tastes change. And if the only thing you know how to do is target current popular taste, then you are in deep weeds when tastes change. Imagine trying to sell a script that does not provide the audience with a satisfying ending. What if the last episode just fades to black? Oh, wait, is that what the Sopranos did? Yeah, it worked well for that show. But, I doubt that many other shows will follow suit.

For those who find the use of templates repugnant for quality stories because it constrains the writers creativity, I would respectfully suggest that the writer can put their creativity into the parts of the story that can benefit from it the most while tweaking existing templates for elements that are less important. No point in reinventing the wheel in different shapes and colors when you can put that energy into new ways to use the wheel.

The point here is that templates provide some amazing benefits while also introducing some significant risks. But, what if there were numerous templates and one could choose the parameters of the story they are trying to write, feed them into a piece of software and let the software generate the script. Let's take this  a step further and say that the software is intelligent and it can figure out what needs to be and then does it. Hmmm.. this is starting to sound like artificial intelligence writing stories. Is that possible? Well, just wait till the next post.

Writing as an Evolving Prototype

 Once you see writing as a design and development activity analogous to writing software, you begin to see how some lessons from software development can be applied to writing stories.

Back in the early days of software development, say the mid to late 1970's, most software development followed a widely accepted process model called the life cycle model or the waterfall model. On paper it was fairly straightforward. The steps in software development were (roughly): requirements, analysis, design, coding, and testing.  Sometimes steps were merged such as in requirements analysis, analysis and design, or coding and testing. The lifecycle would require several years and millions of dollars before the users of the system saw anything concrete and upwards of 75% of the time, they were not happy with what they saw. So, a new idea in software development began to emerge.

Innovative developers began building smaller pieces of the system, showing the smaller pieces to users, and getting feedback much sooner. In hindsight, it seems obvious. But it wasn't. I was involved in several of these projects and the advocates of traditional development models resisted this new approach like the plague. They claimed that all these smaller pieces were throw away pieces and hence a waste of time and money. They advocated building the entire system and then fixing problems in a phase called maintenance. 

Those in favor of building smaller pieces came up with a name for their approach. They called it prototyping. And, in response to critics who called it throwaway software, they started calling it an evolving prototype suggesting it wasn't throw away. It was evolving into the final product. Prototyping became increasingly more popular while lifecycle methods became increasingly more formal. Eventually, developers began to notice that prototyping worked well on certain kinds of software development while more formal methods worked well on other kinds of software development. And they even discovered the key difference. 

On systems that were well understood, the lifecycle models seemed to work best. On systems that were not well understood prototyping seemed to work better. Prototyping is a means of gathering requirements and understanding what the user wants. If you already know that it is much more efficient to just write down what you want and have the developers develop it. Again, all this seems so obvious in hindsight. But it took decades to become obvious.

So, we have two factors to consider here: certainty and efficiency. And we can apply this gem from software development to writing stories. If you know what you want to write and know how to write it, then write down your outline and follow it to write your story. This is commonly known as the plotter approach. If you are not sure what you want to write or are not sure how to write it then you need to try something. This is commonly known as the pantser approach. Or, if we wish to give it a more dignified name, we could call it the evolving prototype approach. 

Thinking about writing fiction as analogous to writing software provides and additional benefit. We wish to write our stories as efficiently as we can given our level of certainty regarding what we are trying to achieve and how to achieve it. In an earlier post, I said "the work of writing is reduced by working out the design to determine the structure rather than by having it emerge through rewrites." This still holds. There may be things we are certain about such as how to write a particular kind of character. While there are other things we are not as certain about such as how to best describe a setting that we have never actually experienced. In this case, you should focus on reducing uncertainty wherever we can. And then crank out the parts where you have a high level of certainty later. The reason why this order is most likely the best is that the reduction of uncertainty in some aspects of writing the story may create uncertainty in other parts.

Another thing we can take from software development is that you should understand the difference between Learning vs Performing. When you have high levels of certainty regarding what you want and how to achieve it, just do it. This is the most efficient way. We call this Performing. When there is uncertainty in what we are trying to achieve, there is stuff we need to learn or figure out before we can proceed. We call this Learning. If you try to Perform when you need to Learn, you will just reduce uncertainty in the least efficient way.

One final word. This is not intended to be set in concrete. You should tinker with it to make it work for you. Each writer is different, and each story is different. Just do what you think is best and if you make a mistake, learn from it.