For example one may chose to come across weak by using the passive voice, either to avoid responsibility, or to pay tribute to another person, or even to create a favorable background for a strong stance on a different subject.
Passive voice itself is not a problem, lack of discretion is.
A couple well known things the field mostly agrees: there are no real synonyms, in the sense that every word has its own semantic baggage, and that manifests different meanings in different contexts. So i.e. "use" and "utilise" are not at all the same words, one fancy.
Same goes for syntactic structures. There's a variety of approaches here, but you'd be hard pressed to find even a supporter of transformational generative grammar, the Chomskyan paradigm who says active and passive sentences come from the same underlying deep structure (roughly speaking), who'd say active and passive voice sentences are equal or equivalent in discourse.
Similarly, you can't boil down "complexity" to the length of sentences or number of clauses. I'd be more willing to concede if this was talking about the distance of things that refer to each other, or the number of words that refer to previous discourse or outer world (i.e. long-distance dependencies & deictic elements). But you can have paragraph long sentences that read just buttery smooth, and "sentence" itself is a pretty vague term that you can't really pin down. Like, if I splat that previous sentence into two around "and", would it really be two sentences, or is it really one sentence to begin with?
All these weird stuff like "don't use active voice" or "use 'simpler' words" etc. are the product of the same mentality that in ye olde times wanted you not to split infinitives, not to end your sentences with prepositions, and other nonsense up with which you should not put.
Trying to talk about it a bit myself, I think I should start with saying that I don't really know methodology in pragmatics, but it and discourse analysis (DA) are pretty close to each other. Discourse analysis is an umbrella term for many research methodologies, and it's a hugely multidisciplinary field, so it's hard to pin it down. Tho suffice to say the concerns of DA inside of linguistics is separate from that in literary studies and the Foucauldian tradition, which tend more towards philosophical approaches. DA in Linguistics is more exact in general and focused on extents of written or aural or signed text and conversations. One of the most common tools is transcriptions peculiar to DA. There on we investigate different properties like structure, pauses or intonation in how they relate to different pragmatic goals, like turn taking in speech and signaling coherence, deixis, etc., in more "purer" DA research, and other strands of research like Critical Discourse Analysis or Feminist Discourse Analysis may then extrapolate how these reflect power relations or social preconceptions. This probably overlaps a lot with pragmatics---and a lot of theoretical and analytical tooling like speech act theory or Gricean maxims are shared---but AFAIU DA is more interested in textual (i.e. speech, not necessarily written) context here, whereas pragmatics in more mechanical and semiotic workings---tho I doubt subfields of linguistics are as distinct as some of the literature makes them seem to be. In any case, a more concrete method that's used a lot in DA (or other subfields) is corpus analysis, where large, often annotated corpora is used in order to test what constituents are found together.
Talking about your interest in math here, you may or may not find what you want in this. On the one hand, language is an incredibly flexible mechanism and almost everything, from words to farts to where one looks at has semiotic and discourse-relevant content, and this is something that conflicts a lot with more logically / mathematically motivated approaches like Chomskyan grammar. Simply put, language is hard to pin down, because unlike say in physics, chemistry, or astronomy, your subject matter is an extremely diverse, constantly changing beast that's produced with animals with extreme agency. But OTOH universe, biology, or materials are messy in their own ways and mathematical approaches have been useful in these pursuits. So it depends on where you come to it from, really.
As to the article you linked I'd say it falls under the umbrella of computational semantics which is an area I'm totally alien to (FWIW I'm more partial to usage based grammar and sociolinguistics). It does use corpora but that's not exclusive to discourse analysis (in fact its a methodology that began around the 90s and is used across all subdisciplines of linguistics today, maybe bar phonetics). There's an over reliance on the concept of the "word" as some basic elementary unit here, which is not really the case. There being a word to express something in a language or not is not really a barrier or obstacle to expressing it in any language. The study feels like it could really make use of another linguist among its authors. There are disciplines that explore language contact and variation, second language education, and translation in itself, which could have a word or two here.
I got some thoughts on that. I'm a student, see.
They call me master, but it don't mean much. Learned a bit from Aronoff down around San Pedro. Rees-Miller too. Real men of discipline.
'bout myself, I got thoughts on discourse, but that's pragmatism talkin'
I hear you like math though. Chomsky's a son of a bitch, and he do know grammar. But he can't put you in a box even if you talk animals, or rocks - 'cause he don't know where you come from.
And that thing you asked me to read?
I read it.
Keeps the rain off, but ain't that what semantics is all about? Keeping concepts from seeping into your cloths to dampen your soul?
Your brain?
I said too much already, but discipline's a creature of words. A silver toung'd preachers words if you ask me.The best community I know of is the LinguistList. I follow their RSS feeds to keep on top of what's happening in the field, but they also maintain many mailing lists, which you might wanna give a look.
There's also linguistics.stackexchange which does have some good content and answers.
There's some small linguist presence on Humanities Commons (https://hcommons.org).
LingBuzz is like an arXiV--discussion board hybrid thing.
If you're into conlanging, I know that some communities/subreddits exist but IDK the names or URLs, unfortunately.
Other than that sadly we don't have anything to the extent that CS or physics enjoy.
Edit: You're welcome, BTW. If you want to ask about anything more specific, resources or otherwise, I'd love to help.
"X can be changed to y" comes across a lot softer than "you should change x to y", and is likely to get better results.
Weak language, such as overuse of adverb qualifiers, is much more akin to the passive-aggressive tone that you’re referring to. Though, it can also just be the product of overly self-conscious writing.
Also, intent is just a small part of context. Therefore, it would rarely if ever supersede the importance of context. For example: Two spies are sitting on a bench speaking in code to one another. Spy 1 gives up an important detail concerning the number of Tridents in a specific location in exchange for some equally important detail related to Spy 2’s national defense structure. Now, it may not have been Spy 1’s intention to divulge his nation’s exact percentage of Tridents stored in a single location, but nonetheless and unbeknownst to Spy 1, Spy 2 already knew the total number of Tridents and therefore could deduce from context. Shovel in a few more cascading deductions and maybe Spy 2 now knows a lot more than Spy 1 ever intended to tell him.
This is true of a lot of things. You can make racists statements out of ignorance without intending to be racist. However, many things that may seem racist on the surface may in greater context be criticism, parody, or just an attempt at candid documentation. All of which are partially informed by intent, but only because intent is a part of context.
Given this and people’s proclivity for hiding their intentions (and even with intentions stated can be either a lie or the truth and we wouldn’t really know so we have to yet still include it in our box of contextual evidence that is still yet to be determined, in terms of authenticity) I would choose to gather as many contextual clues as I could while keeping intent at arms length—still there and I see it, but maybe not the most important factor.
Just for fun the “editor” put this comment at grade 10 and told me to aim lower, grade 9. Too many hard-to-read sentences, adverbs, and difficult words like “therefore”.
At the end of the day, it is still up to the writer which suggestions they keep. Anyone who blindly accepts suggestions because the tool recommended them probably wasn't in the correct mindset to write well anyways.
To that extent it seems like it could be useful in certain contexts.
Exactly. Hemingway himself wasn't a mechanical automaton blindly following his own rules. He broke them when it suited him. Style in writing isn't a process that can be applied in a mechanical fashion. Instead, _taste_ must be exercised.
It is not meant for writing beautiful, stylish prose. Or, at least, that's not what I'd ever consider using it for.
I'm not a proselytizer for it, but this is one reason I appreciate codebases that enforce arbitrary per-line character limits. It pressures me to game the utility:noise ratio.
Left to our own devices, we occasionally bang out 40-char variable names, of which 20 chars add no value to the reader. The color column on the screen is a nice coach, saying "c'mon, you can do better".
I'm always super curious about how someone's method of learning a second language impacts the way they use that language. My casual hypothesis is that what you read while learning impacted how you write, while what you listened to affected how you speak -- do you find that to be true?
Once I heard that writing and talking involve different areas of the brain.
For what it's worth, I read it as "with all its quirks and complexity" because my brain just filled in the "its" I expected, which it might not have done if the English had been worse.
Good writing is not an intrinsic property of the text - it is defined within a context of message and reader. Who is this product for? Business correspondence, tech specs, technical books, or school essays? The style is usually not consistent throughout the same text. Intro or side notes are often less formal than the body.
Applying the same standard of writing to all texts preaches mediocrity as a virtue. I would prefer a tool that instead of telling me what to do, would only offer insights into the text. With AI's insights of sentiment, writing style, and several metrics of complexity, one could use that guidance for the same goals as this app and many more.