TOEIC Link Reading — Part Seven Online-Chat and Text-Message Strings: Tracking Turn-Taking, Speaker Intent, and the "Why Did the Writer Say This" Item

The online-chat and text-message-string passages in TOEIC Link Part Seven are the item set candidates most consistently underestimate. The passage is short, the register is casual, each line is tagged with a sender name and a timestamp, and nothing in the text is technically difficult. Candidates read the thread the way they read a real group chat — fast, for gist, half-attending — and then answer the questions from a vague memory of "what the conversation was about." The item writers build the entire set to punish exactly that reading. They ask why a specific writer sent a specific line, what a one-word reply means, who a pronoun refers to three turns later, and what will happen next. None of those questions can be answered from gist. They can only be answered by reading the thread as a structured exchange of turns, each turn responding to the one before it.

The accuracy pattern proves the point. Candidates who treat chat strings as easy warm-ups score them at roughly the same rate as the hardest triple-passage sets, because the "easy" surface hides intent and inference questions that require the same close reading as a dense article. Candidates who slow down for the six or seven lines of a chat thread, mark who is talking to whom, and read each turn as an answer to the previous turn, convert the same set into three or four near-automatic points. The difference is not reading speed or vocabulary. It is the decision to read a casual thread with the same discipline you would give a formal letter.

Why the chat-string genre is harder than it looks

A chat or text-message string compresses a conversation into fragments. Real speakers share context, so they leave things unsaid — they use pronouns without repeating the antecedent, they answer with a single word, they reference an earlier plan without restating it, and they signal agreement or refusal through tone rather than explicit statement. The passage reproduces all of this compression, and the questions target precisely the parts that were left unsaid. This is the same inference discipline that formal passages demand, applied to a surface that discourages you from applying it. For the underlying skill, work through pragmatic implicature and conventional inference recognition; the chat string is that skill's highest-density testing ground.

There are three compression features that generate almost every hard question in the set.

The first is ellipsis — the omission of words recoverable from context. When one writer asks "Can you cover the Thursday shift?" and another replies "Wish I could," the reply omits everything except the emotional stance. The candidate must reconstruct the full meaning: the writer cannot cover the shift and regrets it. An intent question will ask what the writer means by "Wish I could," and every distractor will offer a literal or reversed reading.

The second is deixis — words whose meaning depends on who is speaking and when. "I'll handle it tomorrow," "send it to me," "that won't work for us" — each pronoun and time word points to a specific person and moment that the candidate must pin down from the sender tags and timestamps. A reference question exploits the fact that "me" refers to a different person in each turn.

The third is adjacency — the way each turn is glued to the turn before it. A reply of "Sure, 3 p.m. works" only makes sense as an answer to a proposed time. If you read the reply in isolation, you lose the proposal it accepts. The whole thread is a chain of these paired turns, and the questions test whether you tracked the chain.

The five recurring item types in a chat string

Every chat-string set is assembled from the same five question molds. Learning to recognize the mold from the stem tells you what to look for before you re-read.

Item type one — the intent question ("Why does the writer say X")

This is the signature item of the genre and the most-missed. The stem quotes a short line — often idiomatic, elliptical, or ironic — and asks what the writer most likely means or why the writer wrote it. The answer is never the literal surface of the quoted line; it is the communicative function of that line in its slot in the thread. To answer, find the quoted line, read the turn immediately before it, and ask what job this line is doing in response — accepting, declining, correcting, reassuring, hinting at a problem. The correct choice paraphrases that job. Distractors offer the literal reading, a plausible-but-unsupported motive, and a function that belongs to a different turn.

Item type two — the reference question ("Who / what does 'it' / 'they' refer to")

The stem asks for the antecedent of a pronoun or a vague noun. In a multi-party thread the antecedent is often several turns back and belongs to a different speaker than the one using the pronoun. Track the sender tags: the referent of "it" is whatever thing was last introduced as a topic, not whatever noun is physically nearest. This is the same discipline as cohesive-device and reference resolution in Part Six, transferred to a multi-speaker surface where the added variable is who is speaking.

Item type three — the detail question ("What does the writer agree to do")

The stem asks for a concrete fact stated somewhere in the thread — a time, a task assignment, a location, a next step. These are the most reliable points in the set because the answer is explicit, but you must locate the exact turn and read whether the action was agreed, refused, or deferred. The trap is a fact that was proposed by one writer and then rejected or modified by another; the passage states the proposal in plain words, and a careless reader marks it as the answer even though the thread overturned it two lines later.

Item type four — the inference question ("What is probably true about the situation")

The stem asks you to combine several turns into a conclusion the thread implies but never states — that the writers are colleagues, that a deadline is near, that a delivery is late. Gather two or three explicit clues and check that the conclusion follows from all of them, not from one. The correct choice is the minimal conclusion the clues support; distractors overreach into details the thread never gives.

Item type five — the next-action question ("What will the writer most likely do next")

The stem asks what happens after the thread ends. The answer is the action the last turn commits to or sets up. Read the final two turns, identify the commitment ("I'll email the file now," "Let me check and get back to you"), and choose the action that directly follows. Distractors offer actions that were discussed earlier but superseded, or actions assigned to the wrong writer.

How to read a chat string in the exam

Reading a chat thread well is a fixed four-step routine that costs about twenty seconds and pays for itself across four questions.

Step one — map the speakers. Before reading for content, note how many people are in the thread and glance at the sender tags. A two-party text exchange behaves differently from a three-or-four-party group chat, where "me," "you," and "they" multiply. Knowing the cast prevents reference errors.

Step two — read turn by turn, pairing each reply to its prompt. Do not skim the block. Read line one, then read line two as an answer to line one, and so on. Mentally label each turn's function: proposes, accepts, declines, asks, clarifies, jokes. The functions, not the words, are what the intent questions test.

Step three — mark the pivots. A pivot is a turn that changes the plan — a refusal, a new time, a problem raised, a decision reached. Pivots are where detail and next-action answers live, and where the "proposed then overturned" traps are set. Note them as you read.

Step four — read the questions and return to the exact turn. With the thread's structure in memory, each stem points you to a specific turn or turn-pair. Re-read that turn, confirm the function or fact, and match it to the choice. Never answer an intent or reference question from memory; the traps are built for readers who do.

The distractor patterns unique to chat strings

Chat-string distractors exploit the genre's surface. Recognizing the pattern lets you eliminate before you deliberate — the same elimination discipline formal sets reward, tuned to this genre's traps.

• The literal reading of an idiom. For "Wish I could" or "Don't hold your breath," a distractor offers the surface meaning while the correct answer captures the implied refusal or warning.
• The reversed stance. For a hedged decline ("I'll see what I can do"), a distractor reads it as a firm yes. The thread's later turns usually disambiguate; read them.
• The superseded fact. A time or task that was proposed and then changed appears as a distractor in its original form. Only the final agreed version is correct.
• The wrong speaker. For reference and next-action questions, a distractor attributes the pronoun or the action to a plausible but incorrect member of a multi-party thread.
• The overreaching inference. For situation questions, a distractor states a conclusion that is likely in real life but unsupported by the specific turns given.

For a fuller catalogue of how paraphrase and trap answers are built across all of Part Seven, cross-train with paraphrase recognition in Part Seven, which teaches you to see the same idea behind different words — the exact skill the intent question demands.

A four-week protocol for the chat-string set

The chat-string set responds quickly to focused practice because the item molds are so few. Four weeks of short daily reps moves most candidates from "I lose one or two of these every test" to "these are automatic."

Week one — turn-function labeling. Take five chat-string passages and, without answering any questions, label the function of every turn: proposes, accepts, declines, asks, clarifies, jokes, warns. The goal is to read a thread as a sequence of moves, not a block of text. Check your labels against how the questions are worded; the intent stems will confirm which functions the writers intended.

Week two — intent questions in isolation. Do fifteen to twenty intent questions ("Why does the writer say X") back to back. For each, write one sentence stating the function of the quoted line before you look at the choices, then pick the choice that matches your sentence. This builds the reflex of deriving the answer before the distractors can bias you.

Week three — reference and next-action drills. Alternate reference questions (pin the antecedent using sender tags) and next-action questions (read the final two turns). These two types share a mechanism — precise attribution to the right speaker — and drilling them together hardens the habit of tracking who is doing what.

Week four — timed full sets with a pivot log. Do complete chat-string sets under time. For each set, before answering, jot the one or two pivot turns. Then answer. Review every miss by asking which of the five distractor patterns caught you, and log it. By the end of the week the pivot log should show the same two or three patterns repeating, and naming them is most of the cure.

Sequence the chat string with the other short Part Seven genres so the transfer compounds: email headers and subject lines for fast context trains the same rapid speaker-and-purpose mapping on single-sender documents, and multiple-passage cross-reference strategies extends turn-tracking into the harder linked-document sets. Together they turn the "easy" end of Part Seven — where careless reading quietly bleeds points — into the most dependable scoring block in the section.

The one habit that fixes the chat string

If you change only one thing, change this: read the casual thread with the discipline you reserve for the formal letter. The chat string's entire difficulty is a disguise. Short lines and a friendly register signal "skim me," and the item writers count on you obeying. The candidate who instead reads six lines as six deliberate moves — each one answering the last, each one doing a job the questions will name — turns the most-underestimated set in Part Seven into free, repeatable points. Slow down for the easy-looking thread, and it stops being where you lose the section.