For older posts, please visit the archived words / myth / ampers & virgule blog.

full view
grid view

AI handles a softball question better than I expected

You have probably been reading about AI and in particular about ChatGPT. And if you’ve been reading critically, you know that ChatGPT is prone to wildly fanciful speculation. So I’ve been avoiding it. But a friend posted something fun on Facebook, along with a signup link, and I said, Hey, what the heck. I signed up.

My friend’s challenge worked as expected, which is to say badly. The challenge was to enter “Please write a short biography (up to 3 paragraphs) for [YOUR NAME], who lives in [YOUR CITY, YOUR STATE].” I did that and learned all sorts of new facts about myself, none of which were true. Go ahead. Try it.

But then I decided to ask it to write a brief essay in response to this: “What is the difference between a book designer and a graphic designer?” [Spoiler: I’ve written about the difference many times in many places. I know the answer.] Here is ChatGPT’s first attempt at an answer:

While there can be overlap in their skills and responsibilities, there are distinct differences between a book designer and a graphic designer.

A book designer specializes in the design and layout of books, focusing on the visual presentation and organization of printed materials. They consider factors such as typography, page layout, cover design, and overall aesthetics to create a cohesive and visually appealing book. Book designers understand the intricacies of book production, including selecting appropriate typefaces, establishing a hierarchy of information, and ensuring readability. They work closely with authors, publishers, and printers to bring the book’s content to life in a visually pleasing and engaging manner.

On the other hand, a graphic designer has a broader scope of work and can be involved in various design disciplines. They may create visual designs for a wide range of media, including branding, marketing materials, websites, logos, packaging, and more. Graphic designers use their artistic skills, knowledge of design principles, and software proficiency to develop visually compelling and communicative designs. They often work on projects that require a strong understanding of branding, user experience, and visual communication across different mediums.

While both book designers and graphic designers share a foundation in design principles and aesthetics, their areas of specialization and the nature of their work differ. Book designers focus primarily on the specific requirements and considerations of book production, whereas graphic designers have a broader scope and can work on various design projects across different industries and mediums.

Not bad for a robot. But I wanted it to be more specific. So I asked this: “What are the risks of having a graphic designer design a book?” The response follows:

While graphic designers possess design skills that can be valuable in book design, there are certain risks involved when using a graphic designer to design a book. Some potential risks include:

  1. Lack of Book-specific Expertise: Graphic designers may not have a deep understanding of the unique considerations and technical aspects specific to book design. Books have their own requirements such as proper pagination, margins, gutter space, and chapter breaks. Without a comprehensive understanding of these elements, a graphic designer might overlook crucial aspects, resulting in a book that is visually inconsistent or difficult to read.
  2. Inadequate Typography Knowledge: Typography plays a crucial role in book design, affecting readability and aesthetics. While graphic designers are typically skilled in working with type, they may not have the specialized knowledge required for book typography. Issues such as appropriate typeface selection, leading, kerning, and ensuring consistency throughout the book can be overlooked, leading to a suboptimal reading experience.
  3. Insufficient Knowledge of Publishing Standards: Book production involves adhering to specific publishing standards and requirements. This includes understanding industry-standard file formats, color modes, image resolutions, and print specifications. If a graphic designer lacks knowledge of these standards, it can result in technical issues or printing problems that affect the quality of the final product.
  4. Limited Understanding of Narrative Flow: Book design involves considering the narrative flow and pacing of the content. It requires a keen eye for structuring chapters, integrating images or illustrations, and maintaining a consistent visual theme throughout the book. Graphic designers without experience in book design may not fully appreciate these considerations, potentially disrupting the reading experience or the intended impact of the book’s content.

To mitigate these risks, it is often advisable to collaborate with a professional book designer or someone with expertise in book design. They possess specialized knowledge and experience specific to book production, ensuring that the design aligns with industry standards, enhances readability, and supports the overall intent of the book.

The voice is flat and academic, but the content is thorough and accurate. On the one hand, this is scary as all get-out. On the other hand, it will be interesting to see how the public responds to the acres of blandness being spilled all over websites and media of all kinds. I hope I don’t find a lot of AI-generated text in manuscripts I edit. That would be disappointing.


This is a brief technical essay for editors. Writers are free to listen in.

Why do we argue, er, debate, er, discuss commas so often? Why do we mostly all agree about one comma and have widely divergent views about another comma? (Okay, there are only three possible views for any given comma: needed, forbidden, and optional; so maybe “widely divergent” is the wrong formulation.) Why do editors insist on the rightness of their view?

Relax. There’s room for more than one approach, and readers adapt smoothly and without confusion. You will be a better editor if you acknowledge that there are different ways to approach comma usage that are equally valid.

Consider: We understand that the ur-comma was a breath mark, pure and simple. It was an aid to the lector, when few were literate and reading was done from the pulpit. But also remember that some modern writing is meant to be read aloud, whether to children or to a theater audience or at a poetry slam.

I recall an interview on NPR with Robert Alter, when he was still working on his translation of Genesis or perhaps when he had just completed it. The phrase that struck me, as he was explaining his frequent use of repetition and reduplication, was that “some people read with their eyes, and some people read with their ears.”

So that’s one axis: Do you use commas strictly analytically (following formal rules), or do you use them to guide the reader in the intended prosody of the sentence? Obviously, you want to have one consistent approach in a given text, and if you only ever edit that same kind of text (academic work in the humanities, say), then you will always use the same approach. But other kinds of texts may benefit from using a different approach.

Now let’s look at the other axis, which extends from close punctuation (close-p) to open punctuation (open-p). Close-p is characterized by more punctuation in general than open-p, more commas and more semicolons, in particular. Open-p is characterized by a minimalist approach to commas, semicolons, and colons. If it is has more of anything, that would be the dash.

I’ve taken a stab (based on pure guesswork, mind you) at placing various sorts of prose on this plane I’ve defined. You might place them differently or think of better examples, and that’s fine.

My main point is that all of these exist in the world, and chaos has not ensued. So saying, as some editors are wont, that style guide A or style guide B is the gospel handed down from above and all other possible approaches are wrong does not seem productive to me.

What makes sense to me is to realize that all of these styles are valid in specific contexts and that to the extent an editor can let go of the idea that one style is better than another in some absolute sense and can instead embrace the idea of choosing a punctuation style that suits the text at hand, that person becomes a better, more versatile editor.

We can still use analysis to describe the reason for using or omitting a particular comma, but we can do it in the context of understanding the range of possible styles.

The lowly speech tag

I think it’s time for a general reconsideration of the convention around commas and speech tags. I’m sensing some grumbling among the ground troops (fellow editors), and I think it may be time for some brave style guide author or three to tackle the problem.

Here’s how I see it. In the sentence “Juanita said ‘I’m coming too,’” the quotation is a clause that is the direct object of the transitive verb said. Simple subject-verb-object (SVO) order, the canonical order in English. If, for variety’s sake, we sometimes switch to OSV order (“I’m coming too,” Juanita said.), we need a comma between the quotation and the speaker. Similarly, if we switch to OVS order (“I’m coming too,” said Juanita.), we need a comma. And if we go all weirdlike and use VSO (Said Juanita, “I’m coming too.”) we need a comma. In all three cases, we need the comma because of the inversion and not because of the presence of a quotation.

Now at some point (I haven’t been able to pin down when this happened, but I want to say post 1900), writers started to reanalyze the situation and came up with the “rule” that there is always a comma between the quotation and the rest of the sentence. So they back-applied this idea to standard SVO sentences and we ended up with “Juanita said, ‘I’m coming too.’“ And that’s the rule we all learned in school. Quite recently (2005 is the first example I can find, but that’s probably not really the first time), we started calling the SV or VS part a speech tag, and now people debate whether something is or isn’t a speech tag and therefore does or doesn’t require a comma. But I think this is just a result of not analyzing the sentence grammatically in the first place.

I think that if we start consistently dropping the comma in SVO constructions, we’ll all be happier, and the style guides will catch up eventually.