On reading, comprehension, and having original thoughts
To read with diligence; not to rest satisfied with a light and superficial knowledge (Meditations, Marcus Aurelius
Reading is my main source for new knowledge. It offers explanations for the surrounding world and simply an enjoyable pastime. But for an embarrassingly long time I didn’t realise that there was a gap between how I read and my goals: Comprehension, Recall, and having Original Thoughts. The knowledge I gathered was often shallow. Sometimes I wasn’t able to reconstruct the reasoning behind a concept, or I wasn’t aware of the limitations of an idea. In other words, I was missing thorough comprehension and the necessary scepticism. To make matters worse, I also forgot most of the acquired knowledge a short while after.
The main reason for my experience is rooted in the reading approach I acquired during my time in university. Here, the incentive was more external and focused on collecting as much information as possible in as little time as possible just to dump it into an exam. What resided in my memory had to vanish to make room for the next chunk of information and the next exam. Rinse and repeat. That process brought me successfully through my studies, but was detrimental to sustainable knowledge acquisition. What I lacked was a principled approach to reading.
This text is the result of an on going process to close the gap between my goals and my reading approach. It is of interest to anyone who experiences the same mismatch or who wants to improve his current reading approach. It describes how to work through books and papers and what type of written works to concentrate on.
- The beginning
- The reading approach
- On reading original works and derivatives
- Final words
The goals I set out before developing a structured reading approach were the following:
For the scope of this text, it describes the ability to recreate an idea or concept from scratch, together with all its reasoning. It is also the knowledge about an idea’s limitations and assumptions.
This reading approach has to support moving information into long-term memory through time-spaced repetition. But since there are no guarantees that all information stays in memory, a compressed version of the book’s content is needed to allow for fast and easy access.
Having original thoughts isn’t meant to be an absolute statement. It is relativistic and personal. Great books can inspire new ideas in a reader and let him have an original thought. They create new connections and extend his understanding. They also offer that satisfactory feeling of having discovered something new. But in the end a reader’s original thought probably isn’t novel at all. Someone, somewhere most likely has thought it already, maybe he even went so far as to write it down.
As luck would have it, I came across Mortimer J. Adler’s How to Read a Book while I started to look into different reading approaches. His process appealed to me immediately. It is simple and structured, with thorough comprehension as a core goal. Mr. Adler wants us to have a conversation, I would even say debate, with the author. We have to follow his arguments and proofs and try to find counterarguments and criticism. From my understanding it should resemble the dialectic method. It should be like one of these long conversations one has with friends till the dark of night. A back and forth of ideas, arguments and counterarguments.
I used Mr. Adler’s approach as a starting point and adjusted and optimised it where needed until I ended up with the first stable version described below.
The reading approach
When I say first stable version it means the process went through many phases and still continues. It is bound to continually evolve to ensure that everything that hampers learning progress can be improved. To describe its shape as volatile also serves as an argument for continued experimentation until the reading it fits like a glove. That is true for me and it should be for the reader who adapts the process to himself.
In general, the process in its current shape is split up into five distinct stages:
Getting an overview of a book’s content, so we are sure it has what we are looking for. It also serves as a way to create a mental map of how that content is structured and conveyed to us.
Thoroughly reading through each chapter. At the same time creating notes and highlights which will form the basis of a condensed representation of the book.
Reiterating a chapter’s content, after some time passed, by working through notes and highlights. Doing that makes sure we only extract the information most relevant to us. The repetition also helps memorization.
Combining your notes and highlights into a new document that describes the overarching concepts and your additional thoughts.
Finally, answering open questions and formulating criticism to extend our understanding and gauge the limits of ideas and concepts expressed in the book.
All these steps are structured in a way to guarantee that we can stop at any time and continue later with a minimal amount of effort. Thereby, we are not forced to commit a significant amount of time upfront. That makes it compatible with other commitments everyone has in life, like family and work.
Stage 1: Get an overview by skimming the book
Before we start our journey through an author’s book we have to make sure it actually is the thing we are looking for. Nothing is more frustrating than to find oneself half way through a book just to realise it is not elaborating on a specific topic or is missing crucial information.
Skimming sets an entry bar for books to pass before we read them, but it also helps to create a mental map of their structure. We know in which order what content is unveiled and where the author tries to lead us. It creates a common theme, a story line we can follow so not to get lost on the way.
To start off we read the introduction, or abstract if it is a scientific paper, and the conclusion if available. Afterwards, we go through each chapter by reading the heading and subheadings and apply a brief study of all figures, charts, and tables. It is already a sign of quality, from my experience, if figures and the like have descriptions that make them self-explanatory. Making highlights or writing notes during this phase should be avoided. The goal is to get a first impression and an overview as fast as possible. In the end, we might decide not to read the book and any extra time investment is wasted.
For a brief moment now, let us concentrate on headings. Sometimes during our studies, we will experience what can only be called creative headings. They do not convey the goal or content of a chapter, but try to be clever and artistic. They are more akin to headings one would find in novels. But they should not be used in works for knowledge acquisition. Creative headings do not help our understanding. We are not able to grasp the meaning until we read the chapter and sometimes even then the meaning might be lost on us. To give a concrete example, take a look at Black Swan. Even though I enjoyed the book, its headings are atrocious. What does “Umberto Eco’s Antilibrary, or how we seek validation” mean? At the same time, experiencing unfavourable headings should make a reader realise and appreciate written works with good headings. To also offer an example here, an interested reader can take a look at The Wealth of Nations or Benjamin Graham’s The Intelligent Investor.
Stage 2: Analysing a chapter
When we have decided to study the book we will read through each chapter thoroughly. The first goal of this step is to identify and understand the author’s ideas. The second is to begin the compilation of a compressed representation of the book itself by using highlights and notes.
We start by reading the content with care and without haste. There is nobody who is hurrying us through a written work; we can and have to invest the proper time. Of course, there can be exceptions to that rule when we for example work through a class assignment. But I exclude that for now and assume an intrinsic motivation to read.
Let us concentrate on the specifics of taking highlights for a moment. As said before, we will use highlights to build up a compressed representation of the book. The question is: What should we highlight? During the time I developed this approach five categories crystallised, which should guide the reader in his decision. Text should be highlighted:
When it seems to be part of the concepts or ideas explained in that chapter. That is, the actual knowledge the author tries to transfer.
When it is a genuinely interesting statement. That might overlap with category (1) or could be something independent.
When it is unknown to us. That is, knowledge which is a requirement. Here additional study is necessary to be able to understand what the author is going for. While reading I usually just skim the topic in question, often using Wikipedia, to get a rough idea. Thus, I am not interrupted from continued reading in the current work for a longer period, which at least to me is unsatisfactory and distracting.
When it looks somehow wrong, as if the author made a mistake and we start to formulate a question or criticism. Here it is, that we develop the deepest understanding of the underlying work and its limitations. It is also the part which can and often will require another significant time investment.
When it is missing supporting facts and proofs. Either we have to find them or it turns out to be a criticism, because of the lack thereof.
The reader is best advised not to highlight all text passages in the same color. Otherwise, he will invest a significant amount of time afterwards in reconstructing the categories again. He might choose a color for each category. As a personal preference, I combine (1) and (2) under the same color.
Of course, sometimes it isn’t enough to just highlight a passage of text. Some statements or even whole works can be extraordinarily dense and requires extra information to be comprehensible. One example of such a book is John Maynard Keynes’ The General Theory of Employment, Interest and Money. Here, writing down notes next to a highlight is a necessity. As a general guide:
We add information which we need in order to understand a text passage, like context or a more verbose version of the passage.
It is advisable to just write down a couple of words and leave the complete formulation to a later stage. That makes sure we are not blocked for long, but the note should still provide enough information so we don’t forget what the highlight was all about.
While reading we will encounter the phenomenon that some criticism, questions or overarching concepts cannot be highlighted directly in the book. There is simply not a single sentence or paragraph expressing it. The only option I found is to write them down somewhere else, for example into an notebook. Again, a personal recommendation is to not fully formulate them at that stage. Just noting some words and phrases and putting references to the book or external resources should be enough for the beginning.
At some point we will reach the end of the chapter. Depending on the length and hardness of it we might either continue reading or we go to the next stage of this reading approach.
Stage 3: Reiterating a chapter by working through highlights and notes
When we decide that there is enough raw material, meaning highlights and notes, we start to iterate it. That means we read through all of the material we haven’t yet processed.
We adjust a highlight’s range if it isn’t capturing all the text necessary to unveil its meaning. That includes extending and reducing the range. Personally, I rather tend to mark a bit more and have a proper sentence. It makes reading them again more pleasant. And we have to keep in mind that highlights are part of the compressed representation of the book, which we will look up and read frequently.
We add or remove whole highlights where needed.
We formulate our notes into proper paragraphs.
Of course, we now have to read up on the unknown concepts from category (3). When we only read the introduction of a Wikipedia article before, we now read it completely and some additional text if necessary to get a better overview. If the topic at hand should warrant a deeper analysis I prefer to put a book into my [Anti-Library]. After we have done all that we can put a concise summary into our notebook. The last step is to remove highlights of this category from our book to not clutter the content we try to extract.
Where we described open questions, criticism and overarching concepts before only with a few notes and references, we now fully formulate them. After some time passed, with continued reading and putting some thought into them, we will reach a point where we are sure that it actually warrants that extra energy and we also have a better understanding of them. So it could happen for example that a former criticism vanishes because we realise a flaw in our understanding.
After we finish post processing our highlights and notes, we compile a compressed representation of the underlying chapter. We also recalled all necessary information, which will help in persisting them in our long term memory. The effect is similar to Spaced repetition. Besides that, going through that material also helps to discover more overarching concepts or criticism. Some thoughts develop over several pages or even sections and sometimes go unnoticed while we are reading intensely.
Producing this extract also makes sure we don’t have to read the book in one go. We can just stop in between and come back later. To be able to continue we only have to take a glance at our highlights and notes to refresh our memories.
Stage 4: Extracting highlights and notes into a stand-alone document
When we reached the end of the book we compiled a collection of highlights and notes, which are scattered across the book itself and a notebook. In this shape all this is scarcely helpful. To refresh ones memory or to search for specific information will be a tedious task. What we actually need is a single place holding all the work we have done.
Creating that representation will be more or less time consuming depending on the readers setup. If he is a digital reader he will probably have an export option in his app and can extract all highlights and notes without much effort. On the other had, if his approach is more traditional it falls upon him to write out all highlights.
The next step of this phase is to create a heading structure for the exported highlights. I recommend discarding the author’s headings and beginning anew by keeping the original chapter structure and by splitting the highlights apart based on topics. By doing so, we are able to partition information into coherent groups and provide headings which are meaningful. Apart from that, it also forces us to read through the content, which helps memorization and might unveil again some hidden criticism, question or knowledge.
When all this work is done, we not only produced a multilayered representation of the knowledge of the book, but hopefully also developed a thorough understanding of the topic at hand.
Stage 5: Working through open questions and criticism
Over time some of our questions and criticisms will come up in our thought process more often than others. When that is the case, we start new investigations, which can range from reading additional books and papers, or do some analysis, or simply to think about the problem longer.
Often it is here that new ideas arise. Maybe there isn’t a ready answer to a question and we have to find it on our own. Maybe some of the criticism we have leads to a change of the underlying knowledge. It is also here, that we as readers are active and continue the story. We left the known path given by the author, his guidance and have to find our own way.
On reading original works and derivatives
There are, of course, many books worth reading well. There is a much larger number that should be only inspected. (How to read a book, Mortimer Adler)
So far we discussed how to accomplish the goals of Comprehension, Recall and having Original Thoughts set out at the beginning. In this chapter we will examine which types of written works are best suited for this set of goals. Depending on the expected level of understanding we can make a distinction between what can be called Foundational Works and Derivatives.
It is the first written work which puts a concept or idea in a coherent form. It is often the foundation of a field of knowledge. Examples are The Wealth of Nations by Adam Smith which is the foundational work for classical economics, or Albert Einstein’s Special and General Theory of Relativity which created the basis of our modern understanding of gravity.
Works which are based on or derived from these foundations are Derivatives. They extend the foundations, provide interpretations and opinions, or produce a partial, often condensed view of the matter. Translations can also fall into this category, but we will discuss this specific topic later on.
Of course, it isn’t a statement about a work’s quality if it falls into one category or the other, but rather helps us to make a decision when to read what kind of book or paper.
Foundational works for thorough understanding
I recommend to an active reader to go with foundational works when he tries to achieve thorough understanding. The reasons for that are given in the following section.
Building a knowledge foundation
These works are the first of their kind. They do not require a prior understanding of the field to be comprehensible. They do not cut away the context and history of how the work came into being. Foundational books and papers provide material to build up a strong knowledge foundation which is a prerequisite to understanding works which extend its field.
Furthermore, when we acquired these basic mental tools we can put derived works into their context. We understand what they try to change and where their limitations are.
Getting an unaltered view
Derivatives may alter the foundation works and therefore the view on its knowledge. That change can happen in different ways.
Influence of the author
The author who creates a derived work might inject his opinion and ideology. He then will put more emphasis on ideas and concepts which will help his cause and neglect others. Another influence on an author’s work is his own comprehension. There is the possibility that he doesn’t understand the topic properly himself and therefore his derivatives have to reflect that lack of understanding. When the foundations aren’t strong how can the building on top be?
The type of work
Even if the author stayed close the original work and understood it properly, the type of work he tries to derive might change the view in itself. If his goal is to produce a summary the reasoning cannot be as thorough, the history and context as complete. On the other hand, if the goal is to extend the field there is usually no point in rephrasing its whole corpus of knowledge again. The author will formulate his findings and describe how it connects to the corpus. It is upon the reader to have the required understanding of the field’s foundations.
The above example describe derivates which alter the foundations to some extend, but there is one type of derived work which tries to be as close to the original as possible: the translation. But even here alterations have to happen. Translating text suffers from the illness of information loss. The meaning of a phrase in one language might be hard to translate into another. The meaning changes, so does the view.
Limitations of reading the originals
Of course, there are practical limitations to a foundational works first approach. Sometimes they simply aren’t the right source. For example we want to get get an overview. A summary might yield better results here. Another hindrance can be language. We decide to read a book written in a foreign language we haven’t mastered. In that case, we only have two options: we learn the language or we use a translation. Which one is more appropriate depends solely on the reader and his preferences. In the end, the goal is understanding but the time investment most be sensible.
One limitation specific to the digital reader is the availability of the work as digital copy. If it doesn’t exist it cannot be integrated into the reader’s setup. Fortunately, for many foundational works the Gutenberg Project did a tremendous job in providing different digital formats at no cost. We also can find prominent books in their digital forms on platforms like Amazon. Where there is a demand, there will be supply.
Derivatives for getting an overview or extending our knowledge
Starting with the foundations makes sure we are building up the same foundations in our head, a prerequisite for continued reading and research. But not every art and science is equally interesting or important us. When only an overview or rough grasp of the concepts of a field is required working through a summary can be an advisable approach.
On the other hand, when we already have a basic understanding we might extend that with new criticism we didn’t find on our own, new ideas and concepts which derive from the originals, or to bring the foundations into a modern context.
What type of written work we select will depend on multiply factors and we have to analyse them before we make our decision. But if our goal should be thorough understanding of a field we are well advised to start with the foundational works, because they will provide us with the necessary basis.
I will use the final words of this text to emphasise two points important to any reader who might decide to adopt the approach described above. First, it has its limitations. This process cannot be applied to every book. For example, the content of some written works is truly dense and using highlights would extract significant parts of the text, which doesn’t make for a compressed representation. More on that in Appendix 3 . This reading approach worked for many historic foundational books and papers, but it can fail with other types. We have to keep that in mind and continue experimenting and improving this reading approach when friction arises.
One last point I want to emphasise is that any interested reader should go back to the original source of this text, Mortimer Adler’s How to Read a Book. He describes, better than I can, why it matters to read and to read thoroughly and his arguments are even more prevalent today as they were during his time.
1: Personal reading setup
In this section I will describe my personal setup for those who are interested. I am a digital reader. I have my books, highlights and notes with me no matter where I go. I don’t need to remind myself to take all the reading necessities with me all the time, they are just there. This setup also makes sure that backups are created frequently.
I use Amazon’s Kindle app across all my devices. It supports highlights and note taking and has an export function. Unfortunately, the app only produces HTML which doesn’t read well in a document. Therefore, I transform it into Markdown which I then embed into my document. The transformation is done with the following script.
For taking external notes and collecting all highlights I use Bear App. It has convenient features like tagging of notes, cross-referencing between notes and synchronising across all my devices.
2: PDFs to ePub and Mobi
Scientific papers are provided as PDFs more often than not. But reading a PDF can be impractical, especially on a mobile device like a phone. Their layout isn’t adjusted for different screen sizes and synchronisation of highlights and notes across devices isn’t a given. To better fit PDFs, scientific or not, into this reading approach I found a relatively straight forward way of transforming them into eBook formats. This runs under the assumption that the reader has not MS Word.
Convert the PDF into a MS Word document. There are website like PDF to Word Converter which provide such a service.
Open this Word document in Apple’s Pages app, adjust the layout if necessary.
Export the result to ePub.
This should provide a reasonably good eBook quality even with scientific papers. In case he has MS Word, it might be enough to do step (1) and export the result directly to a eBook format.
3: Educational and densely written books
There are types of books which hardly fit into our reading approach. One case, where I tried and failed are educational books. Often their content is truly dense. They try to cover a whole field of knowledge in a couple of hundred pages. The reader just has to remember books he had to read for school or university classes. They hold many if not most of the models, concepts, theories and ideas of a field, but often without any context or history. Applying our reading approach would therefore yield a massive representation which would not be significantly smaller than the book itself. This observation can be generalised to any book which has a high knowledge density. A non-educational book which comes to mind is Marcus Aurelius’ Meditations.
For now, my only recommendation is to restrict the process of highlighting to categories (2), (3), (4) and to put a greater emphasis on writing notes into a notebook.
A personal library represents a corpus of books a reader has read so far. It is also a physical, or digital, representation of the different fields of knowledge a reader is knowledgable about. It can also be an indicator of his depth of understanding of a specific field. What a library isn’t representing is the lack of knowledge and understanding, the anti-xknowledge. Based on that realization Nassim Taleb described the concept of an Anti-Library in his book Black Swan. An anti-library not only contains the books read but also books the reader hasn’t gotten his hands on so far and therefore shows knowledge that he lacks. It is an exercise in personal sceptisism, to understand ones own limitations and should act as a humbling experience.