Benchmark: It don't mean nothing. I quit before I stop.

[Bootstrap]

Neto posted an example where Google AI does poorly. The reason it does poorly is this: Google searches for the text on the Internet, fails to find it, and summarizes the web pages that it does find, trying to find a connection to the original text.

There are other reasons that Google fails on this:

1. Logically, "I quit before I stop" doesn't make sense. But it does have meaning.
2. This is a small snippet of text from a broader context, which Google does not have access to. A text without a context is only a pretext.
3. An awful lot of meaning is in the shared common ground that we know, but don't explicitly mention.

I used different tools and approaches and got rather different results than Neto did. But in this post, let me repeat Neto's message:

I was watching an old movie (uncertain of the date, but probably from the 40's or before), and a lady started crying. A man asked her, "What are you cryin' for, Lady?"

She says, I" been crying a lot lately. It don't mean nothing. I quit before I stop."

So I typed that last sentence into Google search. Here's what the AI deal "said":

AI Overview
This phrase means that the speaker will immediately stop doing something ("quit") as soon as they are no longer able to continue ("stop") - essentially indicating an intention to leave or cease participation the moment they encounter a barrier or reach a limit.
Key points:
Directness:
It's a very straightforward way to say that there will be no delay in stopping once a certain point is reached.
No flexibility:
The phrase implies that there is no room for negotiation or trying to push past the stopping point.
Context matters:
Depending on the situation, it could be interpreted as a firm commitment to leave quickly, or a warning that the speaker may not be able to continue much longer.
Example situations:
In a competition:
"I'm going as hard as I can, but I quit as soon as I stop feeling my legs."
In a work setting:
"If the workload becomes too much, I'm going to quit as soon as I stop being able to manage it."

I admit it didn't stand a chance interpreting something like that, but it didn't just say that it's nonsense, or extremely bad grammar, or something of that sort.

EDIT:
I wasn't sure of the movie title, but it says 'Tomorrow". Turns out it was a 1972 black & white movie, with Robert Duvall & Olga Bellin. It's just set back in time, and since it's B & W, I figured it was older.

[Bootstrap]

I have had good luck with the OpenAI 4o LLM for analyzing texts. You can do this yourself by going to chatgpt.com. You may get different results than I do, because I analyze a lot of texts in a variety of languages and it does adapt to the user.

First, I asked it to identify the movie. It could not. So I know that the results are not informed by the rest of the movie. Ideally, I would locate or produce a transcript for the movie, especially if we want to use the result in some serious way.

I started with a good generic prompt. The "non-native speakers of English" is meant to draw out things that we understand as native speakers which are not taught in books and to keep the language of the response simple:

Explain this text to non-native speakers of English: lady started crying. A man asked her, "What are you cryin' for, Lady?"

She says, I" been crying a lot lately. It don't mean nothing. I quit before I stop."

Here's the GPT response:

This text can be challenging for non-native speakers because of its informal language and idiomatic expressions. Here's an explanation:

Story Summary
A woman starts crying. A man notices this and asks her why she's crying. She responds by saying she's been crying a lot recently, but it doesn't mean anything important. Then, she adds a puzzling statement: "I quit before I stop."

Breaking It Down

1. "What are you cryin' for, Lady?"
   • This is informal speech. The word "cryin'" is short for "crying." The man is politely asking why the woman is crying.
   • "For" is sometimes added in conversational English, even though "Why are you crying?" is also correct.
2. "I been crying a lot lately."
   • This is grammatically informal. It means: "I have been crying a lot recently."
   • The speaker uses "I been" instead of "I have been," which is common in casual speech.
3. "It don't mean nothing."
   • This is also informal grammar. It means: "It doesn’t mean anything."
   • Double negatives like "don’t" and "nothing" are often used in informal English to emphasize the idea, but they are not correct in formal grammar.
4. "I quit before I stop."
   • This is a bit of a paradoxical statement. The woman seems to be saying that she gives up on crying ("quits") before she's fully finished crying ("stops").
   • This could mean:
     • She starts crying but doesn't let it fully take over her emotions.
     • She's acknowledging that her crying is frequent and doesn't follow a clear pattern.

Overall Meaning
The woman is emotional and has been crying a lot, but she dismisses its importance, saying it's just something she does lately. Her last statement may reflect her complex feelings or her difficulty controlling her emotions.

This dialogue highlights informal speech, idiomatic language, and emotional expression, which might be tricky for non-native speakers to understand without explanation.

[Bootstrap]

Looking at these results, I felt that it would do better if analyzed the text on the basis of discourse flow and discourse features. Neto knows about these things. This LLM has a different take on discourse features than Runge / Levinsohn, scholars who have done good work for Greek and Hebrew, so let's ask GPT how it understands these things:

What is Discourse Flow?
Discourse flow refers to how ideas, topics, and information progress naturally within a conversation or text. It’s about the logical and cohesive movement from one part of the discussion to the next, making the overall communication smooth and understandable. Key aspects include:

• Transitions: How ideas or speakers move from one to the next.
• Continuity: Maintaining focus on the topic.
• Patterns of emphasis: Highlighting or expanding important points.

For example, in a story, discourse flow might move from setting the scene to introducing a conflict, then resolving it.

What are Discourse Features?
Discourse features are the tools or elements used to structure communication and convey meaning in conversations or texts. These include:

• Turn-taking: How speakers take turns in a conversation.
• Cohesion: Words or phrases (like "and," "however") that link sentences and ideas.
• Topic management: How topics are introduced, maintained, or changed.
• Politeness strategies: Ways to show respect or manage social relationships (e.g., "please").
• Repetition or emphasis: Repeating key points for clarity or importance.

Discourse features help shape the interaction and ensure the message is understood in the intended way.

[Bootstrap]

Next, I asked it to do this:

Use discourse features and discourse analysis

The result was academic and hard for non-specialists to read, so I also asked it to do this:

Take the results of the discourse analysis and retell it in plain English for people with a high school degree

Here's the result:

Here’s a plain English retelling of the discourse analysis:

1. What's Happening
A woman is crying, and a man asks her why. She tells him she's been crying a lot lately, but it doesn’t really mean anything. Then she adds something unusual: "I quit before I stop."

2. How They Talk to Each Other

• The man is curious or concerned, so he asks her a polite question: "What are you crying for, Lady?" He uses casual language, like "cryin’" instead of "crying," which makes the conversation feel informal.
• The woman gives a simple explanation: she cries a lot, but it doesn’t mean much to her. Then she says something that’s hard to understand: "I quit before I stop." It sounds like she’s talking about her emotions, but it’s unclear what she really means.

3. The Woman’s Emotions
Her crying shows she’s upset, but when she talks, she tries to act like it’s not a big deal. She doesn’t explain much, which makes her emotions seem complicated. The way she says "I quit before I stop" might mean:

1. She feels overwhelmed and can’t explain why she’s crying.
2. She often stops herself from fully dealing with her feelings.

4. The Tone of the Conversation
The whole conversation feels casual, almost like two strangers talking briefly. The man uses a friendly tone, and the woman tries to dismiss her feelings as unimportant. Even though they’re talking about her crying, it’s not a deep or serious conversation—at least on the surface.

5. Why It’s Hard to Understand
The woman’s last line, "I quit before I stop," is confusing. It doesn’t follow regular logic, so the man (or the audience) has to guess what she means. This kind of statement makes the conversation more emotional than logical.

6. What This Interaction Shows

• The man seems polite but doesn’t push her to share more.
• The woman shows she’s struggling but doesn’t want to talk deeply about her emotions.
• Their conversation is brief and doesn’t really solve anything, but it gives us a glimpse into the woman’s emotional state.

This retelling simplifies the analysis and focuses on the main ideas, using language that’s easy to follow while keeping the emotional depth of the scene.

[Bootstrap]

The more complete discourse analysis that I started with may be interesting to Neto, it will make most people's eyes glaze over. Maybe Judas also knows this stuff? Who else?

Here it is for anyone who can read it without their eyes glazing over. It does contain some useful insights that were not in the plain English summary.

To explain this text using discourse features and discourse analysis, we can break it down into its conversational elements, social context, and underlying meanings.

1. Context and Participants

• Participants: The conversation involves a man and a woman (lady). The man initiates the dialogue after noticing the woman's behavior.
• Context:
  • Informal and likely spontaneous interaction.
  • The woman is emotional and vulnerable.
  • The man is curious or concerned but remains casual in his approach.

2. Turn-Taking and Interaction

• Turn 1: The man speaks first, taking the role of the initiator.
  • "What are you cryin' for, Lady?"
    • The question is direct but softened with the word "Lady," which personalizes it.
    • The use of "cryin’" signals an informal tone.
    • The man displays an inquisitive stance, seeking clarification about her emotional state.
• Turn 2: The woman responds, providing both an explanation and a reflective comment.
  • "I been crying a lot lately. It don't mean nothing."
    • The first sentence offers a statement of fact, placing emphasis on frequency ("a lot lately").
    • The second sentence downplays the importance of her crying, showing an attempt to dismiss or minimize her emotional display.
  • "I quit before I stop."
    • This adds a complex, reflective statement that disrupts the expected turn structure. It introduces ambiguity, requiring interpretation.

3. Discourse Features

• 3.1. Informality:
  • Use of contracted forms ("cryin’," "don’t") and non-standard grammar ("I been," "It don't mean nothing") reflects casual, spoken English.
  • These features establish an informal and interpersonal tone, aligning with a scenario where emotions are raw and unfiltered.
• 3.2. Emotional Expression:
  • The woman’s crying is a non-verbal discourse marker that frames the interaction.
  • Crying acts as an emotional cue, prompting the man’s question and structuring the discourse as one that centers on feelings.
• 3.3. Ambiguity:
  • "I quit before I stop" is deliberately vague and paradoxical.
    • It resists simple interpretation, encouraging the listener (and reader) to infer deeper meaning.
    • This is an example of indirect speech, which conveys internal conflict rather than a clear message.

4. Cohesion and Coherence

• The interaction is cohesive because the man’s question directly relates to the woman’s visible emotional state.
• Coherence is disrupted slightly by the ambiguous statement ("I quit before I stop"), which requires additional cognitive processing to connect to the rest of the discourse.

5. Power Dynamics and Politeness

• The man’s use of "Lady" shows positive politeness, aiming to soften the directness of his question and convey respect or concern.
• The woman’s response exhibits a mix of negative politeness (minimizing the importance of her emotions to avoid burdening the man) and self-revelation (sharing her frequent crying).

6. Implicit Meaning and Pragmatics

• Pragmatic Markers:
  • "It don’t mean nothing" signals that the woman wants to downplay her emotions, suggesting either embarrassment or a desire not to draw attention to her struggles.
  • "I quit before I stop" implies an unresolved internal conflict. Pragmatically, it could indicate:
    1. Emotional exhaustion.
    2. A pattern of suppressing feelings before they escalate.
• Relevance Theory: The man’s question assumes that her crying has a clear cause, but her response challenges this assumption, emphasizing the unpredictability of emotions.

7. Conclusion
This interaction is a brief yet rich exchange that uses informality, emotional expression, and ambiguity to convey deeper layers of meaning. Through discourse analysis, we see:

• How conversational roles (initiator vs. responder) frame the interaction.
• The role of ambiguity and indirect speech in reflecting complex emotions.
• The interplay of politeness and informality in managing the emotional tone of the dialogue.

Ultimately, the discourse invites the listener to engage not only with the surface-level words but also with the emotional and relational subtext.

[Bootstrap]

I'm done posting my results. Here's my initial take:

1. The default analysis was not very insightful. It wasn't badly wrong, it just didn't say much to help me think about the nuances of what was being said and what was not being said.
2. Using discourse flow and discourse features, I think it did do a good job explaining the conversation. Telling GPT how to analyze it is important to getting results. Trying different analyses and comparing them is also useful.
3. The plain English version kept many of the insights of the more academic version, but lost some of the texture of the conversation.
4. These results were done taking a small snippet of text out of context. That's not a responsible way to use GPT, I would much prefer to work with a transcript of the screenplay and provide more context.
5. These results should be vetted by human beings before anyone takes them seriously. If I were creating commentary material for translators, I would always have biblical scholars check them.

But I think using GPT for a preliminary analysis can give people useful information to help them understand this exchange in the play. I can tell it what kind of analysis to use and see how well it does. I can try out different formats for the output. I can feed it the texts that I want it to focus on.

[Neto]

I'm done posting my results. Here's my initial take:

1. The default analysis was not very insightful. It wasn't badly wrong, it just didn't say much to help me think about the nuances of what was being said and what was not being said.
2. Using discourse flow and discourse features, I think it did do a good job explaining the conversation. Telling GPT how to analyze it is important to getting results. Trying different analyses and comparing them is also useful.
3. The plain English version kept many of the insights of the more academic version, but lost some of the texture of the conversation.
4. These results were done taking a small snippet of text out of context. That's not a responsible way to use GPT, I would much prefer to work with a transcript of the screenplay and provide more context.
5. These results should be vetted by human beings before anyone takes them seriously. If I were creating commentary material for translators, I would always have biblical scholars check them.

But I think using GPT for a preliminary analysis can give people useful information to help them understand this exchange in the play. I can tell it what kind of analysis to use and see how well it does. I can try out different formats for the output. I can feed it the texts that I want it to focus on.

Thanks, Boot. Yes, I actually did that Google search simply because of the incongruence or dissonance involved in the quotation. No, on the bold face of it, it in nonsensical. Yet this is a part of human speech, and I was just curious about how AI would handle something like this.

Statements like this are also a common joke form. Some examples from an older man in the congregation back home - he was a joker, and loved to "catch people up".
You ask him "How you doin'?", and he would say something like "I feel more like I do now than I did yesterday."
Shake his hand, and he might ask you, "Are you afraid?" Then when you say "No,", he answers "Then why are you shaking?"

These things reflect the complex nature of discourse analysis, and of human conversation in general. In the case of jokes, the context includes a lot of elements that are not reflected at all in what is actually said. The Banawa love to joke. I'll try to give some examples, and make them 'make sense' for the non-Banawa person.
A person comes over because of a pain in their chest. They say "When I breathe, there is a spear in my chest, right here." So I say, "Then don't breathe." Everyone laughs, and then we go on to diagnose the health problem.
Plays on words: They were somewhat puzzled the first time I pulled this one on them, because it's really a linguistic joke. But it only took them a moment to understand, and then they found it very humorous, and repeated it to others. The phrase that means "she said" is identical to a sentence final word that means 'almost". As in "The tree is almost falling over." So, I asked them, Did you hear its voice?" This was a complete nonsense reply, but they caught on to the linguistic similarity between these two completely unrelated things. That is the human element. Can AI be taught to understand this sort of thing? I wasn't really intending my post yesterday to be a criticism - it was a curiosity on my part, to test its 'humanness'. I did, however, expect it to point out the nonsensical element of the woman's statement. You got better results - it at least then pointed out the puzzling nature of her statement.