Two Kinds of Continuity Work: Tracking as You Go vs. Investigating Later

When I first began designing Scriptri, I assumed most continuity work would happen while a scene was being written.

A character learns a secret. A relationship changes. A weapon is introduced. An injury occurs. An important object changes hands.

The writer notices the change and records it while the scene is still fresh.

That is one useful kind of continuity work. I think of it as proactive tracking: preserving important changes as the manuscript develops.

But while working on Scriptri and using it with my own fiction, I realized that this only addresses part of the problem.

Continuity questions often appear later, after a detail has already passed quietly through several scenes:

• When did Elena’s sleeve become torn?
• Did Marcus already know about the ledger?
• Was the North Annex locked before the confrontation?
• Where was the blue key last seen?
• Is this the same mug that appeared earlier, or merely another mug?
• Which character currently possesses the damaged transmitter?

I could not have reliably predicted that every one of those details would matter when I first wrote them.

As I noticed those reactive questions, I began to see a gap in Scriptri’s functionality: sometimes the task is no longer to record a change as I move forward. It is to investigate the manuscript and reconstruct what happened.

That is the second kind of continuity work: reactive investigation.

A useful continuity system should support both.

Proactive Tracking: Record What Changes

Proactive tracking happens when I recognize that a scene has changed something later scenes may need to preserve.

After writing or revising a scene, I might record that:

• A character learns something.
• A relationship becomes warmer or more guarded.
• An object is lost, damaged, or transferred.
• A location changes condition or control.
• A new clue appears.
• A character forms a mistaken belief.
• A deadline is established.
• A question needs to be resolved during revision.

These notes do not need to summarize the entire scene.

Suppose Elena discovers that Marcus lied about visiting the capital. The scene note might say:

Elena now knows Marcus lied about traveling to the capital. She does not know where he actually went or why he concealed it.

That note preserves the change that later scenes need to respect.

Before the discovery, Elena may trust Marcus’s account. Afterward, her dialogue and decisions can reflect her new suspicion.

Proactive tracking works particularly well for changes that are obvious while writing:

• A reveal
• An injury
• A betrayal
• An object changing hands
• A shift in political control
• A promise or deadline
• A resource being depleted
• A relationship turn

The writer already understands that something important happened. The task is simply to preserve it.

The Limits of Tracking as You Go

The problem is that I do not always know which details will become important.

Stories change during drafting.

A background object becomes symbolic. A minor character becomes central to the plot. A casual comment creates a contradiction after a later reveal is added. A detail that seemed decorative becomes evidence in a mystery.

Even when I recognize that something matters as a continuity concern, I may not know which part of it I will need later.

Suppose I write that Elena catches her sleeve on a damaged doorway during an escape. At the time, the torn fabric is only physical description.

Twenty scenes later, another character identifies her from a strip of cloth left behind.

Now I need to know:

• When was the sleeve torn?
• Which sleeve was it?
• Was she still wearing the same clothing later?
• Could the cloth plausibly have been left at the relevant location?
• Did any intervening scene imply that she changed clothes or repaired it?

The detail became continuity-relevant only after the later scene existed.

No amount of discipline can make a writer perfectly predict the future significance of every sentence without excessive planning.

A system that depends entirely on proactive tracking therefore creates an impossible standard: identify and record everything that might later become relevant.

Reactive Investigation: Recover What Became Important

Reactive continuity work almost always begins with a question in some form or another.

The manuscript already exists, at least in part. Something no longer fits, or a detail has become important, and I need to trace its history.

The process may look like this:

1. Define the question.
2. Search for possible references.
3. Review the surrounding scenes.
4. Separate relevant matches from false positives.
5. Identify when the important fact first became true.
6. Determine whether later scenes preserved it consistently.
7. Record the notes so the investigation does not need to be repeated and can be referenced again in the future.

For example:

Where was the blue key last seen?

A search for “blue key” may find several useful passages. It may also miss references such as:

• The key
• The small brass key
• Elena’s key
• The object in her pocket
• It fell down the grate

Search can locate possible matches. The writer still has to interpret what those words refer to.

The reverse problem also occurs. Searching for “key” may return:

• Other physical keys
• A keyboard key
• A map legend
• The phrase “key evidence”
• A metaphorical “key to the mystery”

Search results provide evidence for a reactive investigation. They do not resolve the question by themselves.

Why Ordinary Search Is Not Always Enough

Text search is extremely useful. It is often the first thing I reach for when investigating a fact I need from my manuscript. But ordinary search does not understand identity or narrative context.

It can find every occurrence of “Marcus,” but it may not find:

• Mr. Vale
• The captain
• Her brother
• A nickname
• A title
• An indirect reference

It may also find the same name when it refers to another person.

The same problem applies to objects and locations. Two mugs are not necessarily the same mug. A place may have changed names. A character may use an alias. A damaged sword and its replacement may share the same description.

Search answers: Where does this text appear?

Continuity investigation asks: Which of these passages refers to the same person, object, place, event, or underlying question?

That second question is what I am trying to support with the Investigations workspace: a place to organize search results, preserve relevant evidence, and record the writer’s interpretation rather than treating every textual match as an answer.

Preserve the Investigation, Not Only the Answer

One frustration I have encountered is solving the same question more than once.

The process looks like this: I search the manuscript or my notes, read several documents, reach a conclusion, and continue drafting. Later, the issue resurfaces and I no longer remember which passages I checked or why I reached the earlier answer.

Recording only the final conclusion can help:

The blue key remains in Elena’s coat pocket after chapter 12.

But preserving some of the evidence in a durable record is more useful:

• Elena takes the key in scene 8.
• She places it in her coat pocket in scene 9.
• Marcus searches her bag in scene 11 but not her coat.
• No later scene shows the key being removed.
• Conclusion: Elena still possesses it at the start of scene 13.

That record explains how the answer was reached.

If the manuscript changes later, I can revisit the evidence instead of beginning the investigation from nothing.

This is especially important when the conclusion is uncertain:

The manuscript never clearly establishes whether Elena recovered the key after the fall. Add clarification.

That is still a useful result. An investigation does not have to prove that continuity is already resolved. In fact, it may expose a gap that needs revision.

Some Questions Begin as Continuity Flags

Not every problem needs to be investigated immediately. For example, while drafting, I may notice something uncertain and create a short flag:

Confirm whether Marcus knows about the ledger before this conversation.

Determine when the North Annex was locked.

Check who possesses the blue key after the escape.

That preserves the question without forcing me to interrupt the current scene for a lengthy search.

Later, during revision, the flag can become a focused investigation.

This creates a useful bridge between proactive and reactive work:

1. Notice a possible concern.
2. Record the question.
3. Continue drafting.
4. Investigate when the timing is appropriate.
5. Preserve the conclusion or revise the manuscript.

The takeaway is this: you do not need to solve every problem the instant it appears. Often you won’t even know when a fact is important until you reflect on it later.

A Practical Investigation Method

When a continuity question appears, I find it helpful to phrase it narrowly.

A broad question such as:

What is going on with the key?

is difficult to resolve.

A narrower question is easier:

Who possesses the blue key at the beginning of chapter 18?

Or:

Has Marcus seen the blue key before Elena shows it to him in chapter 18?

Then identify the likely sources of evidence:

• Scenes containing the object
• Notes attached to the object or relevant characters
• Timeline events
• Location records
• Earlier continuity flags
• Mentions under aliases or alternate descriptions

As each result is reviewed, classify it:

• Relevant
• Not relevant
• Possibly relevant
• A different object or entity

Then record a conclusion:

Elena takes the blue key in chapter 9 and still has it in chapter 18. Marcus has not seen it before the reveal.

Or:

The manuscript does not establish where the key goes after chapter 12. Add a transfer or recovery scene before chapter 18.

The second conclusion is not a failed investigation. It has identified the exact revision needed.

Turn Discoveries Into Reusable Continuity

The most useful part of reactive investigation is converting the result into something you can use in your project, later.

Let’s revisit the example of the investigation into Elena’s torn sleeve. Say we discover that it happens during the escape in scene 14.

I can then add a scene-bound note:

Elena tears the right sleeve of her coat while climbing through the damaged doorway.

From that point forward, the detail is no longer buried only in prose. It has become part of the continuity record.

A later scene can account for the torn coat, a change of clothing, or the fabric left behind.

This creates a cycle:

1. A question appears.
2. The manuscript is investigated.
3. The writer reaches a conclusion.
4. The conclusion becomes structured continuity information.
5. Later scenes can reuse that information.

Why Both Workflows Matter

As I build Scriptri, I increasingly find that a continuity system focused only on proactive tracking can feel demanding or only marginally useful. Supporting both workflows is necessary if Scriptri is going to help writers preserve changes they notice immediately and investigate questions that emerge later.

A proactive system asks you to decide during drafting which details deserve permanent notes. That can make every scene feel like a data-entry task.

A system that supports only search has the opposite weakness. It can help recover information later, but the writer may have to repeat the same investigation every time the question returns. The answer never becomes part of the project’s remembered state.

The stronger model combines both:

• Record obvious changes while they are fresh.
• Flag uncertain questions without stopping the draft.
• Investigate details when they become important.
• Preserve the evidence and conclusion.
• Convert useful discoveries into scene-bound notes or other structured records.

The goal is for the project to become easier to understand as you work with it, with relevant state surfaced in the context where it matters.

How Investigations Currently Work in Scriptri

Scriptri now includes an Investigations workspace for reactive continuity questions. It is still an early, in-development feature that will be expanded as I continue to work on Scriptri.

An investigation can begin with a title and a specific question, such as:

Where was the blue key last seen?

It can also be connected to relevant entities or anchored to a scene.

From there, Scriptri can search across several kinds of project information:

• Manuscript scenes
• Scene-bound notes
• Timeline events
• Location events

Potential findings can be attached to the investigation and classified as:

• Relevant
• Not relevant
• Unclear
• A different entity

That last category matters when a textual match refers to something with the same or a similar name but is not actually part of the thread being investigated.

The investigation can remain open, be deferred, or be resolved with a conclusion. This preserves not only the final answer, but the question and the evidence used to reach it.

Scriptri does not decide which search results are correct. It does not infer the truth of the manuscript or automatically resolve ambiguity. You remain responsible for interpreting the evidence and deciding what it means.

You Remain the Source of Truth

This is an important design principle for Scriptri. The software can search, compare, organize, and surface evidence. It can remember which results were useful and which referred to something else. But it does not know the story better than you.

A contradiction may be an error. It may also be a lie, a mistaken belief, an unreliable narrator, an intentional mystery, or a detail awaiting explanation. Only you can decide which interpretation is correct.

The tool’s job is to make the evidence visible enough that the decision does not depend entirely on memory.

Continuity Work Happens at Different Times

In summary: some continuity details surface themselves immediately. A character learns the truth. A bridge collapses. A relationship changes. A weapon is lost. Those changes can be recorded while the scene is fresh.

Other details reveal their importance later. A minor object becomes evidence. A description creates a contradiction. A question resurfaces after thirty scenes. A revision changes the meaning of an earlier passage.

Those details have to be investigated. Both are normal parts of working with a growing manuscript.

Continuity is not only the information I successfully predicted would matter. It also includes the questions I discover after the story has become too large and interconnected to reconstruct casually from memory.

Scriptri is being built to support both: the details recorded as the story develops and the questions that only become visible later.