Thanks to my work at the University of Würzburg I have been able to attend the last Digital Humanities (DH) Conference in Krakow and I am looking forward to be also at the DH Conference in Montreal this summer. The conference in Krakow was one of the best experiences in my academical career. I have been involved with the DH Conference in different roles: participant, co-author, first author and reviewer. Although my experiences is only of two years, I have thought about writing a post about the whole experience, specially for the people that would like to propose for the next year in Ciudad de México. I have already seen many people getting very irritated or even feeling humiliated about the process of proposing. I think the call for papers is only useful for those who already know how the DH Conference works, so let me give you my opinion about some aspects. I am working with quantitative methods to analyse texts, so possibly some things of this post are only true or more accurate for this specific area of DH.
Exercise your Proposal
To present at the DH Conference is like preparing yourself for a marathon: you don’t train some weeks, it takes months, actually more than a year. The call is normally published at the end of the summer. Way before that, you have to know what and with which data you are going to propose. The deadline is somewhere in October-November. In December-January you will receive the comments of the reviewers and you will have the opportunity to answer. In March-April you will know if you are accepted and have to present the final proposal. If you were lucky, you will present in August.
The DH Conference doesn’t work like other humanities Conferences
Let us be honest: the standard way for proposing something to a conference in the Humanities is to write a couple of paragraphs about what you are going to do. It is very possible that you have only done about 10% of what you are suggesting your proposal is going to be about. And that is OK. Normally you don’t even have to add any bibliography. Nobody will complain, specially if the people organizing the section know you. I think we can agree that these are not best practices, but it is how it works.
That definitely won’t work at the DH Conference. And the rest of the DH Conferences (DHd, HDH…) are following this new model.
I cannot tell what works because I have seen how some very good proposals are completely rejected and have attended to lousy long papers. I still have some useful ideas:
- Read all the similar papers both about the technology and similar data that have been presented at the conference in the last 5 years and cite them
- Check for other papers published in other sources about similar data and technology. And cite them
- Define very well your question: Don’t be either general nor greedy, get straight to the point
- Define very well your data: What is it? How big? Why these? Where did you get it? Is it available? In which form and format? More details?
- Define very well your method and parameters: Why these? Are they the best that you could use? Are the scripts available?
- Show results and, if you can, evaluate them
- And try to put all this in way more shorter texts than you would actually need to clarify everything
Beware of the critic!
Prepare yourself to receive the hardest critic about your work that you have heard. Ever. DH is using a way to criticise which is unknown in the Humanities until now. Probably it has to do with the fact that the peer review process is (single) blind, so we are just using the possibility of criticising sincerely the works of our peers without having to assume personal consequences. Maybe we still have to learn how to do so in a more constructive way.
And by the way, you will receive this critic about Christmas time. This year we received the reviewers’ comments exactly on 23th December. Prepare yourself psycologically because right at the beginning of your holidays some anonymous can destroy your work of months. Your holidays can become a “Oh oh oh, Miserable Christmas and Lousy New Year!”.
Using New Technology is not Enough
Let us imagine that you want to propose something about Topic Modelling. No one in your country is using Topic Modelling in your field. You manage to create a big and nice corpus. You manage to use the scripts. You manage to split, preprocess, lemmatize and filter the texts. You manage to make sense out of the output of the Topic Modelling. You make cool visualisations with it. And you write a nice article about it, citing other people’s work. At the end you are exhausted but happy because nobody in your field could have written this. Nice try. Only, it will be most probably rejected at the DH Conference.
Depressing? I know… Unfair? It fills like it. Is that the way we want our field to work? I don’t know… And this is not only truth for Topic Modelling: stylometry, graphs and networks, clustering…
The current expectation of DH Conference is that you have to bring something new also in the method. I think there are three ways how we can achieve that. First and most common, we can get deeper or broader knowledge about already-known-technologies: evaluate its performance, test parameters, apply technology to other kind of data… (example). Second, you can apply to data from the humanities a method not known in the DH, developed in other fields like Computational Studies (example). And third, and that is the most difficult one, you can develop a new way of analysing your data (example).
Everything below that, will tend to be accepted only as a poster. And if. That includes tools and sets of data (like corpora). One example: stylo was a poster in 2011.
A Part of DH has Become Empirical Humanities (and Nobody has Warned you!)
The usual steps in some areas of the DH could be: we go to workshops, to Summer Schools, to pre-conference workshops and we learn about formats, about data, about parameters, about programming languages, about input and output data, about visualisations, about cooler visualisations, and about extra coolest interactive visualisations… You use what you have learned and write a nice proposal. And suddenly the reviewers are all about standard deviations, p values, null hypothesis, confidence intervals, error bars, sampling mean, baselines and significance tests. And you don’t even understand what these things are.
All these concepts are not standard in the great majority of fields of the Humanities, with very few exceptions (like Computational Linguistics). This can be very basic in other fields like Computational Studies or Statistics, but at the end Digital Humanities is about humanists going digital and not about hard-science-people going cultural. Let me rephrase it: the great majority of the people working in the Humanities (Digital or Not-So-Digital) neither use this statistical concepts nor know them. These are also not taught in the great majority of courses about Humanities (Digital or Not-So-Digital). I understand that if we are assuming formats and technologies coming from other fields, it is reasonable to assume also their way of testing results (although we could also apply our traditional humanities-academical discourse to these new digital data). But until the community has developed its ways to become more familiar with them, we should also relax our expectations about them.
In the lasts two years I have been struggling with statistics. The most useful sources for me have been:
- Basic Statistics Course at Coursera
- Straightforward Statistics for Behavioral Science, by James D. Evans
I was going to write: attend one DH Conference before proposing. But since they are all around the world, it might be difficult. Right now the ADHO has the policy of a three year rotation: one year in Europe (2016: Krakow; 2019: Utrecht), one in North America (2017: Vancouver), one in the rest of the world (2018: Ciudad de México). So there are very good chances that the next conferences are going to be expensive for you. It is anyway very helpful to assist one conference to understand better what is expected from the different kind of proposals or how posters work.
Another way to get experience in your proposal is to have a co-author that has already been to one DH conference. Actually since the system of peer-review is only one side blinded, I think that the proposals, with at least one co-author who is known in the typical DH circles, has better chances to get accepted. That is why I think it would be much better to apply a double-blind peer review.
Be brave! It is an interesting Experience
The worst thing that can happen is that you get a mean critic and rejected. So be it. Even in that case, you will learn and get an experience that you didn’t have before. Surely it can help you to write better proposals in the future. It is not the end of the world, nor the end of your career. It is not even the end of this idea, that you can improve, rewrite and send it maybe to a better place in other publications, conferences or even to next DH Conference in a different form. DH Conference is, at the end, one conference among others.
PS: Thanks to the no-blind peer reviewers of this post, who have helped me improve this post with their soft critic