DA Module 12: Processing Speech and Writing
⚡️ Cognitive demands, structural affordances, and a verified AI-supported literature map
Speech and writing place different demands on producers and interpreters, but the contrast is shaped by genre, planning, technology, and context. Listeners normally process a transient signal in real time, supported by prosody and interaction. Readers usually control pace and can revisit a stable text, yet written discourse may package information densely and assume unfamiliar background knowledge. This week begins Unit IV by comparing cognitive and structural affordances rather than repeating a simple speech–writing binary. It also introduces Level 5 AI use: an advanced scholarly workflow. You may use an approved tool to create a preliminary literature map, but every claim, source, and relationship must be independently verified. The documentation should make your orchestration visible, including errors, exclusions, limitations, and the boundaries of what the assigned sources actually support.
Essential understanding(s)
How do the cognitive and structural affordances of speech and writing shape comprehension?
What distinguishes an advanced, transparent AI-supported literature workflow from automated summary collection?
Guiding Questions
How do transience, permanence, pace, interaction, prosody, density, and visual organization affect processing?
How do genre and digital technologies complicate speech–writing comparisons?
What relationships among concepts and sources belong in a literature map?
How can AI-generated claims and references be verified, corrected, or rejected?
Learning outcomes
Compare cognitive and structural demands of spoken and written discourse using authentic examples.
Explain how processing conditions interact with genre, proficiency, background knowledge, and task.
Construct a preliminary literature map with an approved GenAI tool.
Verify every included source and claim against assigned or credible publications.
Document prompts, verification decisions, limitations, and independent synthesis.
Key concepts and vocabulary
| Term | Working definition |
|---|---|
| Transience | The tendency of spoken signals to disappear as they unfold unless recorded or repeated. |
| Recursiveness | The ability to revisit, reread, and reorganize attention within a stable text. |
| Processing load | The cognitive demands imposed by decoding, holding, integrating, and interpreting information. |
| Working memory | The limited system used to maintain and manipulate information during comprehension. |
| Prosody | Vocal patterning that segments, emphasizes, and contextualizes spoken discourse. |
| Orthographic organization | Visual resources of writing such as spacing, punctuation, paragraphs, and headings. |
| Information density | The amount and complexity of meaning packaged within a stretch of discourse. |
| Literature map | A structured representation of relationships among concepts, claims, methods, and sources. |
| Source verification | Independent confirmation that a cited work exists and supports the attributed claim. |
| Scholarly orchestration | Purposeful coordination of tools, sources, judgment, and documentation in a research workflow. |
Main chapter content
-
Processing spoken discourse
Listeners normally receive speech incrementally. The signal unfolds over time, and earlier wording may no longer be available unless it is repeated or recorded. Comprehension requires rapid segmentation of the sound stream, lexical access, syntactic interpretation, integration with context, and updating of a developing mental representation.
Prosody supports this work. Stress highlights information; intonation signals boundaries, stance, and action; rhythm helps listeners group words into units. Visible gesture and shared surroundings may reduce the need for explicit reference. Interaction also distributes processing: listeners can request clarification, and speakers can monitor responses and reformulate.
Difficulty increases when speech is fast, unfamiliar, acoustically degraded, densely informational, or disconnected from background knowledge. Yet spontaneous speech may also be easier than formal writing because clauses are shorter, repetition is common, and meaning is collaboratively repaired. The analyst should relate processing demands to the actual genre and task. -
Processing written discourse
Writing usually provides a stable visual record. Readers can control pace, reread, preview headings, and inspect relationships across paragraphs. Writers can revise and use typography, punctuation, and layout to structure attention. These affordances support complex organization and extended reflection.
At the same time, written discourse can be lexically dense and syntactically compressed. Nominalization, embedded clauses, technical vocabulary, and implicit disciplinary assumptions increase processing load. Readers must often infer the writer’s stance and intended structure without immediate interaction. Visual permanence does not guarantee accessibility.
Reading is strategic. Skilled readers shift between local decoding and global interpretation, adjust pace, use headings and discourse markers, and monitor comprehension. Learners may decode every word yet miss the macrostructure, or predict the general meaning while overlooking critical lexical detail. Unit IV will examine how discourse organization can support these interacting processes. -
Avoiding technological and genre determinism
Digital communication destabilizes simple contrasts. Recorded lectures are replayable. Live collaborative documents are written but interactive. Voice messages are spoken yet persistent. A scripted speech may be more grammatically dense than an informal message. Analysts should therefore distinguish medium from planning, permanence, interactivity, and genre.
Cognitive claims should also be qualified. It is not enough to state that speech is easy or writing is difficult. Processing depends on language proficiency, familiarity, topic knowledge, purpose, attention, accessibility, and design. Captions may support one listener and divide attention for another. Headings can aid comprehension when they accurately represent the structure.
Comparative organizers should use dimensions and evidence. For each sample, describe what the recipient can revisit, which cues segment information, how interaction is possible, and where density or inference creates load. This produces a more defensible account than a list of universal differences. -
Building and verifying a literature map
A literature map organizes a research area around relationships rather than isolated summaries. Nodes may represent concepts, theoretical claims, empirical findings, methods, genres, or pedagogical implications. Links should state the relationship: supports, contrasts with, extends, applies to, or raises a limitation.
At Level 5, an approved GenAI tool may propose categories, search terms, or possible relationships. It may also generate fabricated references or attribute a claim to the wrong source. Every citation must therefore be located independently. Verify author, title, year, publication, and the exact claim supported. If the assigned chapter does not address a proposed point, remove or clearly label it as outside the source.
The record should include rejected nodes and reasons. Perhaps a source was unverifiable, too general, or unrelated to the course question. This negative evidence demonstrates scholarly control. The final map should reflect your synthesis of verified materials, not the architecture of the first AI response. -
From scholarship to instructional judgment
The purpose of the map is to support later instructional design. A claim about transient processing may suggest pre-listening support or chunked replay. A claim about macrostructure may support heading prediction or graphic organization. However, pedagogical recommendations must not be treated as automatic consequences of cognitive theory.
Consider the learner group, text, objective, and setting. A support that reduces load can also remove a desired challenge. Providing a transcript before a listening task changes the task into reading-supported listening. That may be appropriate, but it should be deliberate.
Your Week 12 submission should separate source-derived claims, your synthesis, AI suggestions, and pedagogical inferences. Use assigned readings as the anchor. The literature map becomes credible when another reader can follow each relationship back to a verified source and see how you moved from evidence to interpretation. -
Literature-map verification protocol
Work through the preliminary map one node at a time. For every source, confirm bibliographic existence in the assigned reading, library record, publisher page, or another credible scholarly index. Then open the source and locate the relevant section. Record a short paraphrase in your own words and note the page or chapter location when available. Do not copy an AI summary into the verified column.
Classify each relationship as confirmed, partially supported, contradicted, outside scope, or unverifiable. Partial support is common: a source may discuss permanence but not the pedagogical conclusion the model attached to it. Separate the source claim from your inference with different labels or visual styles.
After verification, reorganize the map around the course question rather than the order of discovery. A useful structure might connect processing conditions to textual features, learner variables, and instructional implications. Write a synthesis paragraph that identifies patterns and tensions across sources. The map should make scholarly reasoning visible; it is not a collection of colorful bubbles or automated abstracts. -
Comparing processing claims across sources
Literature maps become more useful when they represent disagreement and scope. One source may emphasize the transient nature of speech, while another highlights how recording technologies make spoken discourse revisitable. A third may show that genre and task predict processing difficulty more strongly than mode alone. Do not flatten these claims into a single speech-versus-writing table.
Create a comparison row for each major claim: phenomenon, population or context, type of evidence, conclusion, limitation, and pedagogical relevance. This helps distinguish theoretical generalization from findings tied to a specific task. It also prevents an AI-generated summary from treating all sources as equally broad.
When moving to instruction, label the move as an inference. Research describing working-memory demands does not automatically prescribe one classroom technique. Your pedagogical recommendation should explain how the source, learner profile, text, and objective connect. This explicit bridge is central to advanced scholarly orchestration.