Building a citation bank while studying

Dear Scrivener community,

I have a question: I would like to use Scrivener to study by tagging some parts of text with keywords so that I can later on easily search those bits and pieces by keywords and use them as verbatim citations in my PhD thesis, extracting themed collections from different sources. In other words, a bank of (potential) citations.
So far the only method I’ve found is to chop the text at the citation level, thus creating a standalone text, label it at “Citation” and associate such keywords to it.
This method though has 3 drawbacks:
a) I end up with a conspicuous number of fragments for each text I treat; solution so far: group them in a folder;
b) apparently, I can’t perform and advance search specifying a set of keywords AND a specific label [“citation”] (this could be easily worked around by always using the keyword “citation” besides the other keywords, and then do a boolean search, but isn’t there a cleaner way?)
b) it can’t be done “natively” with PDFs, unless I cut and paste the potential citation in another document and I fall back on the standard procedure described above. Again, I can group the original PDF article and the selected excerpts in a folder, to keep the binder tidier, but it seems quite cruncky to me.

I had also thought of using comments while studying the PDF in MacOS’ Preview, but I’ve noticed that comments are not considered/visible when you then import the annotated PDF into Scrivener.

I thought about copying and pasting the citation in the Notes field but then again, the search engine embedded in Scrivener doesn’t allow to search for different type of items (notes, keywords, labels) combined: am I wrong?

I’ve read about tagging software here and there and MultiMarkDown, so I was wondering if you experts could suggest a better workflow the the one I’ve just described.

The radical alternative might be to turn to highly sophisticated (and Windows-only) qualitative data analysis software like Atlas.ti, but I would prefer to keep everything in the same place and in a Mac environment if I can.

Thanks in advance for your precious advice.

You might be better off using a text database (DevonThink Pro) and reference manager (EndNote) for these purposes. They have very sophisticated search abilities, and provide the exact citation forms that you need for footnotes or bibliography.

Personally, I don’t use the Research folder in Scrivener except for short pieces, like a talk or brief essay. Keeping data in separate applications allows me to use them for many kinds of projects. In your career, you are likely to write far more than a dissertation.

I highly recommend the awkwardly titled e-book “Take Control of Getting Started with DEVONthink 2”*, it will cover the various ways you might use DT to keep up with and file away your citations.

  • This title bugs the hell out of me. Why not “Take Control of Devonthink 2: Getting Started”, since the “Take control of” bit seems to be part of a series title. I don’t need to take control of getting started with preparing for learning about wrapping my mind around digging into DevonThink 2. :neutral_face:

Thank you for your pointers.
I get the impression that my English has failed me: by citations I didn’t mean bibliographic references, but verbatim quotes.
I’ll have a look at DevonThink.

Do you have a link to the e-book?

I second (third?) the recommendation of DevonThink. Much better suited than Scrivener for managing massive amounts of research.


I endorse DevonThink Pro as a master reference-bank, from which selections can be withdrawn and placed in Scrivener’s Research folder.

I also agree about the title of the ebook, which is nonetheless worth its (quite low) price. And, whilst on the subject, here it is.

Edit: Beaten to the draw!

Hugh, Katherine,
could you expand a bit on what you find DT outperforming Scrivener when doing research in your cases, with regards to the general question I’ve asked above?
I’ve watched al the video tutorials and played around with it a bit and apart from the fact that allows you to do combined searches (i.e. tags and labels and…) I don’t see much difference from Scrivener. I haven’t read the ebook yet, I have to admit.

Hugh, do you use DT as a sort of catch-all literature database for any interesting document you may want to do a research paper on in the future and then you extrapolate the docs that you think you would need for a specific topic and you bring them to Scrivener research folder? How do you do that exactly?
Is it a sort of enhanced Finder, so to speak?
Because in my case I use Zotero as catch-all database for collecting docs, pdfs and webshots, tag them, add quick notes to them and even cross-refs if needed, and on top of that I have the references as well to be embedded in Scrivener (using the RTF scan feature), thus so far I miss the comparative advantage of DT.


In part answer to your question – the key set of features that distinguishes Devonthink from just about any other information manager that I know of on Windows or OS X is what the developers call AI – “artificial intelligence”. This is coupled with an ability to swallow an immense amount of information without choking: I believe some users have tens of thousands of documents and millions of words in their database. I’ve never used Zotero, but I don’t think it can do this.

AI is a tad hype-ish as a label. But what it means is that the software indexes the words in your documents, so that when you search for, say, “Christopher Columbus”, the software will of course find documents with the explorer’s name in them, but it will also register that, say, the word “Genoa” appears in such documents and may go on to suggest others that contain the name of the Italian sea-port even if Christopher Columbus is not mentioned in them, all ranked in order of relevance. This may or may not be helpful to you; users say that it turns up important stuff that they’ve forgotten they have, and have no idea is relevant, but is; serendipitous connections, in other words. That is also my experience. (Couple DevonThink with DevonAgent and you can do the same kind of thing on the Internet, scouring websites down to several levels.)

Similarly, the software may classify and file your documents without your assistance. So, yes, I 'd use it as a catch-all basket from which I’d take the documents I need for the task in hand (which is what I actually do for fiction, but that’s another story…)

DevonThink has many other bells and whistles, but that is the core of its functionality. Scrivener, for all its virtues, can’t so easily carry as much research and doesn’t have “AI”.

I don’t write non-fiction, but if I did what I would do with DevonThink and Scrivener is, first, plan the structure of my writing in Scrivener with headings and sub-folders in the “Draft” folder, second, put a duplicate structure in the Research folder, third, drag or copy the relevant research documents from DevonThink into the relevant sub-folders in the Research folder in Scrivener, and finally, draft in the two-pane editor with my research for each section or chapter alongside my writing. (This isn’t my scheme: it’s adapted from something the historian Steven Berlin Johnson has written.)

Obviously, for your use of DevonThink and a scheme like this to be worthwhile, you need to have lots of research to store and several research-heavy documents to write.