The Pyramid Principle

The key skill is to recognise which are you major and which are your minor ideas, and to work out their relationships and structure.

The demands of logic and the limitations of readers indicate that it should be of a pyramid shape,

If a person’s writing is unclear, it is most likely because the ordering of the ideas conflicts with the capability of the reader to process. The capability is more or less the same for everyone.

A writer who works to match the flow of the writing to the of his reader’s mind will also clarify his thinking in the process.

The mind already sorts things into pyramids to comprehend them better. If they are pre-sorted already, they are easier to understand. The mind sees things occurring together as belonging together, and imposes a logical pattern.

The pattern will always be a pyramid because it is the only form that meets your mind’s need

to stop at the magical number seven and state the logic of the relationship.

Most minds can remember up to 4 things in RAM When the mind sees the number of items with which it is being presented rise > 5, it starts to group them into categories to retain.

You need to state the logic of the relationship, to be able to move from 10+ items to 3-4, which are easier to remember. Thinking one level of abstraction higher, but the items below it are suggested, because the relationship is also not contrived.

You should present from the top down:

Category > item in it

But you should think from the bottom up:

Items+++ > categorise

When you bring together sentences in a paragraph, there must be a relationship there. They were all needed to express the single idea of the paragraph, which is effectively a summary of them. But once we state the summary sentence, we remove the need for the reader to remember all the sentences, and now they have 1 thing to remember as the main idea.

You keep summarising, paragraphs > sections > on and on, and the document can really only have 1 main idea, one single thought.

Rules for building pyramids:

Ideas at any level in the pyramid must always be summaries of the ideas grouped below them: the major activity in thinking and writing is abstracting to create a new idea out of the ideas below Ideas in each grouping must always be the same kind of idea: you want to state the idea directly implied by the logic of the grouping, so the ideas must all fall into same category. Can you label the ideas with a plural noun? Ideas in each grouping must always be logically ordered: there must be a specific reason why the second idea comes second. There are only 4 possible logical ways in which to order a set of ideas: Deductively (major premise, minor premise, conclusion) - argument order Chronologically (first, second, third) - cause-and-effect relationships Structurally (Boston, New York, Washington) - commenting on an existing structure Comparatively (first most important, second most important) - categorising

The order reflects the analytical process used.

Key to clear writing is to slot the ideas into the pyramidal form and test them against the rules before you begin to write.

You can’t expect to just be ready to write until you’ve symoblized the ideas. The pyramid dictates a rigid set of substructures that can serve to speed the discovery. They are:

The vertical relationship between points and subpoints The horizontal relationship within a set of subpoints The narrative flow of the introduction

The Vertical Relationship

A way to establish a question/answer dialogue with your reader. Captures the reader’s attention.

An idea is a statement that raises a question in a reader’s mind, because you’re telling them something they don’t know. This in turn raises a question in the reader’s head - Why? How?

Then you answer on the line below. However, in the answer there will be other things that the reader doesn’t know, raising further questions.

The writer writes until he thinks that the reader ran out of questions; the reader may disagree, but he will still be able to follow the reasoning logically all the way to the end.

DO NOT RAISE QUESTIONS THAT YOU’RE NOT READY TO ANSWER YET (Disagree: keeping people curious is essentially for capturing their attention). DO NOT ANSWER QUESTIONS BEFORE RAISING THEM (i.e. “Assumptions” area in a document - the reader doesn’t yet have the context when they first see it)

A great value of the pyramid structure is it forces visual recognition of this vertical relationship on you.

The pyramid forces you to present info only as the reader needs it.

The Horizontal Relationship

The points you use to answer the question must be logically structure (inductive/deductive argument - one or the other, but not both at once!)

Deductive grouping: arguments in successive steps. Idea #1 = a statement about a situation in the world. Idea #2 comments on the subject or the predicate of Idea #1, Idea #3 states the implication of the 2 situations.

 E.g.:

 I am a scuba diver. All scuba divers love sharks. Therefore, I love sharks. (I love sharks because I'm a scuba diver.)

Inductive grouping: set of ideas you're able to describe by the same plural noun

 E.g.:

 I love asparagus, avocado, onions, garlic. (I love vegetables.)

If your answer is deductive you know you must have an argument in which the second point comments on the subject/predicate of the first, and the third point draws a 'therefore' from them.

If your answer is inductive you know the ideas must be logically alike and be designated a plural noun.

The Introductory Flow

Must know that the question you're trying to answer has already occurred in the readers mind. If the reader doesn't know something, but also doesn't need to know it, then you have no interest - no communication.

Intro flow comes in as the intro of the history of the origin of the question you're trying to answer.

It's a narrative of events, and should follow the classic pattern of development.

Begin with time and place of a Situation. In that Situation something will have occurred (Complication) causing to raise a Question, to which you provide the Answer.

Knowing the 3 elements of pyramid help:

Vertical - determine the ideas that will answer the questions in the reader's mind Horizontal - ideas you bring together are of a like kind Intro flow - know the reader's question will ensure that ideas are relevant

How to build a pyramid structure

I. The top-down view - things you're most sure of (subject and the reader's knowledge of it)

Use the Intro flow to pull the right points out of your head.

Fill In The Top Box What Subject are you discussing? Draw a box with the Subject in it What Question are you answering in the reader's mind about the Subject? To whom are you writing and what question do you want to have answered in his mind about the Subject? What is the Answer? If you know it

Match the Answer to the Introduction What is the Situation? Prove that you have the clearest Q & A. Take the Subject, move up to the Situation and make the first non controversial statement about it. What's the #1 thing you can say that the reader will agree with? (historically true, knows it, easily checkable) What is the Complication? "So what?" What happened in the Situation to raise the Question? Something went wrong? Do the Question and Answer still follow? The statement of the Complication should raise the Question you wrote down. If it doesn't, change the Question to the one it does raise. Could be the wrong Complication or the wrong Question.

Find the key line What New Question is raised by the Answer? Will you answer it deductively or inductively? If inductively, what is your plural noun?

Structure support points Repeat the question/answer process at this level.

II. The bottom-up approach - when you haven't developed your thinking yet.

Move to the Key Line points and work from there. (NOT VERY CLEAR and POOR EXAMPLE)

If you can't, follow this process:

List all the points you think you want to make Work out the relationships between them Draw conclusions

If your thinking is not clear to the reader in the first 30 seconds of reading, you should rewrite!

Headings are additional communication devices. Use them to help navigate the general thrust of the argument. Don't use "Conclusion", "Findings", etc.

Caveats

Always try top-down (I) first: Once you type it you fall in love with it (100%), so don't just "get it all out". Use the Situation as the starting point for thinking through the intro: Once you know the Situation, Complication, Question and Answer, you can play around with their order, depending on the desired effect. Don't omit to think through the intro: If you have the main point fully stated in your head and it's Question obvious, don't jump to the Key Line items and begin answering the new Question. You'll end up structuring stuff that should be in the Situation/Complication, forcing self into a unwieldy deductive argument. Sort out the intro first, the concentrate on ideas at lower levels. Always put historical chronology in the intro: The body can contain only ideas which relate logically. If you talk about events, then you're spelling out cause-and-effect relationship, not simple historical occurrences that logically don't belong there. Limit the intro to what the reader will agree is true: If the point can be easily checked and deemed true, then the reader "knows" it, even if they didn't know at the start. Support all Key Line points: Answer all the questions raised by the idea. At the key line level, all ideas must have at least 1 level of support. Very true of the "therefore" point in deduction. If there's no need to support the final point, then the argument is over structured and you need to use induction.

Fine points of introduction

Intro is the key step. Summary of what the reader knows = establish the relevance of the question to which you give an answer.

Finding the structure can be complex and time-consuming.

Initial intro is always a story. The reader may not be ready to hear your stuff, amid all the other pressing issues. To captivate him, you need a story.

Every good story has a beginning (Situation), middle (Complication) and end (Answer). But tell the reader a story he already knows or can check if it's true. That way you get him agreeing with you before he reads the rest of the argument.

The intro should be long enough to ensure that you're standing in the same place before diving into arguments. 2-3 paragraphs are enough.

The Situation and Complication can be as long as 3-4 paragraphs. Not more - the person already knows the story. The closer you are in dealing with the person, the shorter the intro. (E.g. ref to the email the person sent you). Just make sure that the reader recalls the Question she asked.

Situation

Make a statement about the subject that you know the reader will agree with because you are telling him something he already knows.

If you struggle, then you've confused about the Subject or starting in the wrong place.

If you're writing for someone identifiable by name, then the situation should just be self-sufficient and non controversial. If it's a book or a blog post, then the question must be planted. Assume that your readers are well informed and present explanation of what is generally accepted knowledge. Arrange known material in a narrative - in a way the reader haven't thought of before - and inspire them the ask the question you wish to address.

Key characteristic = leave the reader expectant of further info.

Complication

Complication to the story.

Described an alteration to a stable situation - not a problem, necessarily (although could be).

 E.g.

 C: Something went wrong > Q: What do we do?
 C: Something could go wrong > Q: How can we prevent it?
 C: Something changed > Q: What should we do?
 C: Something could change > Q: How should we react?
 C: Here's what you might expect to find in it > Q: Do we find it?
 C: Here's someone with a different point of view > Q: Who is right?
 C: In this situation we have 3 alternatives > Q: Which one should we take?

The narrative flow lends a feeling of plausibility to the interpretation, which is helpful in persuading the reader. It gives a sense of rightness to the logic of the writing. It also shows the writer as considerate - someone who aims to be understood clearly.

Intros are meant to remind rather than inform They should always contain the three elements of the story The length of the intro depends on the needs of the reader and the demands of the subject

Common Patterns

Directives Requests for funds "How To" documents Letters of Proposal Progress Reviews

Directives

Telling someone to do something

Planting the question in the readers mind

S = We want to do X C = We need you to do Y Q = (How do I do Y?) [implied - not stated]

If the question is "How", the answer is inevitably "steps".

Request for funds

The question is always "Should I approve the request?"

S = We need this amount of $ C = (We need you to approve it/give it) [implied - not stated] Q = (Should I approve?) [implied - not stated]

"How To" documents

Someone has a problem and you're telling them how to solve it

The structure is always 'steps'.

The intro varies ~ (1) never done before or (2) not doing properly.

(1)

S = Must do X C = Not setup to do Q = How do we get set up?

(2)

The trick is to lay out the process as it's done, then lay out how it should be done

S = Your present state is X C = It doesn't work Q = How to make it work?

Letters of proposal

S = You have a problem C = You have decided to bring in X Q = (Is X the right one to solve the problem?)

The answer should be 'yes' and then the rest explains why:

We understand the problem Here's the approach Here are the people

Most of the work is done in the approach section.

Progress reviews

The formal communication with the client at the end of each phase, leading up to the final report. After the first one, the structure is always the same.

The first one

S = You have problem X C = We told you (in our Letter of proposal) that we would do Y. We have now done Y. Q = (What did you find?)

Once this is done, the client will have a reaction. In the next communication, you refer to it:

S = In our last review we told you X C = You said do Y. We have now done what you asked. Q = (What did you find?)

The pivot on which the doc depends is the Question, of which there is only one to a document. You can work backwards from the key line points, which would help understand what question you are answering with them, which in turn will lead to the complication and the situation.

Transitions

Referencing backward - picking up a word or phrase/main idea of the preceding portion of the pyramid, and using it in the opening sentence.

This is good in the beginning of new chapters, sections or subsections.

Summarizing

If the chapter or section is long or complicated, stop and summarize completely.

Meant to restate the principal matter and tone of the preceding text. Make the job of the reader as easy as possible.

Concluding

Sometimes there's a need to conclude the writing.

This is the place to impose some poeticism, to inspire the reader to want to act. Produce an emotion in them. Give some indication of what he to think about or is able to do with the new knowledge.

Be subtle. Put the significance of your message into a perspective.

Difficult to do well!

Sometimes you might want to do a Next Steps section as a conclusion, but beware of putting things there that the reader might question. The points in conclusion must be obvious logically.

Deduction & Induction

Clear writing results from a clear exposition of the exact relationships between a group of ideas on the same subject.

Deduction forms a line of reasoning that leads to a 'therefore' conclusion. The point one level above in the pyramid is the summary of the line of reasoning, resting heavily on the final point.

Induction defines a group of facts as the same kind of thing, then makes a statement about the sameness.

Deduction

It is the most common way, and easier way than induction. Good way of thinking - bad way of writing.

Deductive argument needs to do 3 things:

Make a statement about a situation in the world Make another statement about a related situation in the same world. The second one is related to the first if it comments on either the subject or the predicate. State the implication of these two situations

It's ok and even necessary to string several arguments together, otherwise the writing comes out as very pedantic.

When to use?

On the Key Line level, do not use it. By going through each step, you're forcing the reader to map the argument entities across different questions.

Instead, it would be best to not drag the reader through the whole argument, but present the message inductively:

You must do X -> How? A1 (B1; C1) > Why? A2, A3 (B2, B3; C2, C3)

Instead of answering the Why > the How, you go How > Why. Then the findings at the bottom level can be formulated into a Recommendations list.

In writing to recommend action, you will never give findings that do not lead to conclusions, nor state conclusions that are not based on findings.

It is better to tell the reader what he should change and then how to go about it, or why. Action before argument = better.

The argument can come first if: The reader is expected to disagree strongly, so you guide him through it first The reader can't understand the action without prior explanation

Push deductive reasoning as low in the pyramid as possible, to limit intervening information to the minimum. At the paragraph level, deductive arguments are lovely and present and easy-to-follow flow.

Inductive reasoning is always easier to absorb at higher levels.

Induction

Much more difficult to do well. More creative activity.

The mind notices that some things are related, brings them together in a group, and comments on their similarity.

To do inductive arguments well, one must learn to:

Define the the ideas in the grouping Find the outliers among them

How it works?

Find one word that describes the ideas in the grouping. Always a plural noun. Check your reasoning, by questioning from the bottom-up.

If the points below completely point to the point above, then you've got it. With inductive ideas, you generally hold the subject constant and vary the predicate, or vice versa.

Highlighting The Structure

The format of the headings must be applied to properly reflect the levels of abstraction in the document. You're going for "ease of comprehension".

Headings

Should represent a division of thought.

Never use only one of any element Show parallel ideas in parallel form: including verb-verb matching and other grammar Limit the wording to the essence of the thought: headings reminds, don't dominate Don't regard headings as part of text: cannot depend on them to carry the message. The document must be able to read smoothly without the headings Introduce each group of headings: state the major point that the grouping is explains or defends Don't overdo: the point is to make it easier on the reader to comprehend

LOGIC IN THINKING

You’re unlikely to produce the first draft that is perfectly logical.

Have a checklist of rules to apply to every grouping.

You can question the general order of the ideas in the grouping You can question the particular source in your problem-solving process You can question your summary statement about them You can question the prose you express them

Questioning The Order

Ideas must be placed in logical order. In deductive argument, this is easy, but in inductive, you have to know how to make the choice.

The ideas are never brought out by chance, they are the result of your analytical process. Only 3 type of analytical activities:

It can determine the causes of an effect: you have determined what effect you want to achieve, and then the action required to achieve it. When there are several actions, there is a system, a process. The grouping of such actions goes in time order because they can only be carried out one by one. It can divide a whole in parts: Org charts/structure of an industry. This is structural order. It can classify like things: Classifying a group as having a distinct quality, like a cluster or several clusters of problems that a company has, for example. This is ranking order.

Time Order

Most pervasive, simple to understand. Spell out the steps a person must take to achieve a particular effect.

Frequent flaw - improper time groupings. Ask yourself, “What would I do first if I were doing this? What second?” Will help uncover where thinking is incomplete, logic confused, grouping false.

Structural Order

What you see once you visualised something.

The pieces of the whole must follow the MECE principle (McKinsey), which means that they have to be Mutually Exclusive and Collectively exhaustive in terms of the whole.

ME = No repetition, DRY CE = What you see is all there is

There are 3 ways to divide:

• by activity (then use time order)
• by the location (structural order, i.e. geography)
• by sets of activities directed at particular product/market/customer (ranking order)

Forcing to see an order if it’s something new helps check that you have been CE for your purposes.

If it’s not an org, then you’re likely analyzing by function. Show the parts in order in which they would be expected to perform the function.

One way to describe a structure is to follow from top down and from left to right. Or you can impose a process order, as on a map. What the reader sees first, second and last is how you describe them. Visualising a process in relationship to a structure is a common device, particularly if you are writing to recommend changes to an existing structure.

Use structural order to help sort faulty logic in a grouping.

Ranking Order

Order or importance. Imposed on a grouping when it brings together a set of things classified as alike, possessing a characteristic in common.

When you think about the “top 3 things”, you already decide that these things are ME = not belonging to the 97 non-top things, and thus also CE because in that class they are the only 3 (top class). Then you rank them strongest-to-weakest.

Ideas may be found to be alike by virtue of one characteristic, but often people use this grouping when the things were produced by 1 process or used in the same structure. Perfectly fine if you are honest about the origin of the likeness (came from the same process - not possessing of the same qualities).

To check the thinking for completeness, you must be able to identify its source and make sure you included all that the source produces.

Why does it matter?

Because the inference you draw determines whether the groupings will be accepted as true by the reader. If over-complete, the statement will not apply equally to the points. If incomplete, then someone else adding more members might draw a different conclusion.

So the process is:

Check that the writing comes from 1 source, and maybe there is a way to split it into subgroups On every level, do the ideas follow in the right order, according to time, ranking or structure

Problem-Solving Process Problem-setting

What is the problem? What does the solution look like? You recognize that a particular solution yields a specific result.

The problems is that (a) you don’t like the result (sales are down) or (b) you cannot explain the result (what determines what we think?)

a) Routine problem solving b) Creative problem solving

For Routine p.s., Sequential Analysis steps:

What is the problem? Where does it lie? Why does it exist? What could we do about it? What should we do about it?

It may not be feasible to do the analysis in such a way due to natural external factors, like people, the source of data, time, etc.

But when you sit down to write about it, try to impose an analytical structure on the findings and the conclusions.

What is the problem?

Create a clear image of what you mean. Some visualisation between good and bad.

Where does it lie?

The set of activities that produce the unfortunate result.

Create a useful structure of the situation, and investigate its parts. When visualising, you’re trying to show how the different elements work together to produce the results. May not be obvious. Diagrams could help.

Otherwise, keep digging until (1) you are sure you identified all parts, (2) you can arrange them in a sequential order, (3) you can clearly show inputs and outputs.

Why does the problem exist?

Process all possible options and elements, but be selective. This is where techniques like PCA or sensitivity analysis could come handy?

What could we do about it?

List all the possible solutions

What should we do about it?

Visualise the new situation with each possible solution involved. Helps find out other points of friction.

What are the risks of each? Incorrect assumptions Not achieve objective Retaliation?

Before you can advise, define clearly (1) the gap between where it is and where it needs to be, (2) the structure of the situation that gave rise to the gap, (3) the structure of the underlying processes, (4) the alternative ways to structure could be changed, (5) the changes required to accommodate the alternative.

Defining The Problem

Where you are now Where you want to be The difference

Use Logical Trees

Financial structure: a diagram based on the financial elements of the system Task structure: the tasks that the company does to perform well Activity structure: the activities that may have led to the result Choice structure: bifurcate trees of choices at each step of the process Sequential structure: sophisticated choice structure

MECE everywhere

Diagrams/mindmaps

An issue is a question so phrased as to require a yes-or-no answer.

Summary Statement

There is a need to state that "this has X problems", then list them. In reality, this is not a clear summary, and if the writer considered the summary, he'd see that it's just summarising the kind of idea to be discussed. Cannot serve as nuggets to focus future thinking. Instead either say, "this has problems", then list the problems.

First, check the origin of the grouping to be MECE: order reflects process, structure or classification Second, look at the kind of statement you're making: either an action (do) or a situation (be, about)

Action

Steps, recommendations, objectives, changes

Use: Writing manuals, action plans, systems

Place each under the effect they're meant to achieve in a causal structure. It helps the reader figure out the process.

This is difficult!

Sorting by effect is alien to most people Separating causes from effects leaves people in chaos Because: Writers state actions vaguely No understanding of cause and effect Want to group by similarity as the most common

Ways to Deal

Make the wording specific

You're able to say about each grouping, "I do these X things to achieve the above effect, then the next X things to achieve the next above effect". Each point must be MECE. The effect must be so specifically stated that it implies an end product. Not "improve profits" because a 10% and 200% would require different steps.

Not always a numerical goal, but always a tangible way to judge completion.

Questions to ask

Does the same step appear more than once? Can I visualise someone taking the action? Will the substeps bring about the step above them? Have I kept the subject the same?

Distinguish the levels of action

Telling cause and effect apart can be tricky. Sometimes actions are worded as conclusions. A conclusion about the cause of a problem implies the action required to correct it.

Importantly, you can't group action ideas by similarity rather than by effect. Otherwise, the grouping will no longer be MECE.

Draw an inference from conclusions

What you can group by similarity are "situation ideas": statements that can be described as "reasons, problems, conclusions.

When you say that the company has X problems, what you did was take a bifurcate division of X problems vs. all the other problems.

Then the statement that defines them will have to be as specific as possible, so as that only X problems can fit this description.

The technique to use is: (1) find the structural similarities in the sentences; (2) visualise the relationships implied between the parts that are similar.

If the subject's the same, you look for similarity between actions or objects. If the actions or objects are the same, look for similarity among subjects.

Sometimes they may seem to be no common ground - which means that the grouping is wrong, and so is the thinking. Re-think!

The impetus to think forward is the reason for drawing inferences. If it doesn't push the thinking upward, there is no guidance forward on the subject.

You can’t group statements together and hope that the reader will understand their significance.

Always ask of any group of ideas, “Why have I brought together these particular ideas and no others?” They either will fall in the same narrowly defined category and be the only in the category - stating their sameness; or they are all the actions that must be taken together to achieve an effect, stating the effect.

Put it in words Visualise the images you used in thinking up your ideas originally.

The trick is to find the nouns and look for relationships between them.

Problem-Solving in Structureless Situations

Sometimes the problems are not of trying to change what a thing is, but to understand what a thing is/should be.

The technique is similar, but requires more visual thinking.

Analytical abduction

You deal with 3 distinct entities:

A rule (belief about the way the world is structured) A case (observed fact that exists in the world) A result (expected occurrence, given the application of the Rule in the Case)

You’re doing Deduction if, Rule > Case > Result: If A then B > A > Therefore, B You’re doing Induction if, Case > Result > Rule: A > B > If A then probably You’re doing Abduction if, Result > Rule > Case: B > If A then B > Possibly A

Analytical problem-solving consists of seeing the Result, looking for its cause in our knowledge of the structure of the situation (Rule) and testing whether we have found it (Case).

In scientific problem-solving, the Rule is unknown. In reasoning to the third, must follow the process:

Hypothesise a structure that could explain the result Devise and experiment that will confirm or exclude the hypothesis Carry out the experiment Recycle the procedure, making sub hypotheses or sequential hypotheses to define the possibilities that remain

Generating hypotheses: drawn out of the air, but suggested by studying the structure of the situation that produced the problem. The design always arrives as a visual image Devising experiments: suggest experiments that will confirm or deny it. It’s important to design it so that it shows a clear “yes or no” answer. Not good enough to just “see what happens”.

Each step demands a clear end product, one you can see. Each image indicates a direction where the analysis should lead.

Solving the problem - representing it so as to make the solution transparent.