December 18, 2010
Use the Active Voice! No, the passive voice should be used! This is a debate that keeps flickering up in the blogosphere.
It taps into two different controversies:
- Subjectivity versus objectivity
- Quality of scientific writing
The subjectivity versus objectivity debate is illustrated quite nicely by the comment thread to Sylvia McLain’s recent post on active versus passive. The central assumption seems to be that using the active voice acknowledges the researcher as an agent that introduces subjectivity into the process. Of course, using the active voice does no such thing. It is the methods that determine the extent to which there is leeway for subjectivity during data collection. If the results leave room for uncertainty or different interpretations (which is almost always the case), this should be made clear in the discussion. Whether the active or the passive voice is used for reporting the work is mainly a matter of convention. While the active voice certainly makes it easier to weave a good yarn, it is perfectly possible to write a bone-dry treatise drenched in positivism and realism without using a single passive. It is also perfectly possible to write a paper replete with passives that is a real rollercoaster of a read. Just take the simple sentence: “These results were not predicted by the standard model of particle physics.” BOOM!
This sentence neatly leads me to my next point. It is often asserted that the active voice is the key to good writing. Well, no, it isn’t. Our sample sentence would not be half as effective, had it been written in the active. As it stands, in the passive, the whole sentence is a single crescendo. We start with the results, then learn that they were not predicted, and then learn that the foundations of physics have just been shaken to their very core. The key information is in sentence-final position, where it is more likely to be remembered due to the recency effect. In English, the sentence-final position is also a classic place for presenting new or salient information. Since English word order is relatively fixed, the passive is often the best solution for moving the agent of an event to the end of a sentence.
The passive also comes in handy when it quite simply does not matter who did what. For example, when I wrote “this should be made clear”, the passive neatly avoided the clunky “author(s)” or the agony of deciding whether to say writer(s) instead of “authors” because it is possible to have authorship without having written a single line. I would imagine that something similar holds when describing the methodology of a standard experiment in biology, physics, and chemistry. (Come on, why say “We added chemical X to chemical Y” when the royal plural is meaningless because the lowly PhD student did all the hard work?) In linguistics and psychology, it’s somewhat different, because we tend to ask our guinea pigs / long-suffering students / valued participants to take in stimuli and do stuff, and this usually lends itself well to descriptions using the active voice.
But if using the active voice isn’t the holy grail of good writing, what is? Well, if you want to know what makes a good text, ask a friendly linguist. Psycholinguists and cognitive psychologists have spent decades studying the effect of linguistic form on the comprehension and retention of read text. There is a whole field of linguistics, text linguistics, devoted to studying the structure of texts. Based on all this research, your friendly linguist would tell you that although active versus passive voice may have an effect on comprehension, that effect is dwarfed by an issue that is often overlooked by grammar fiends: Coherence. A text that does not cohere is not a text, it is a jumble of words and sentences.
There are many ways of achieving coherence. For example, the text could be structured according to an easily recognisable schema, like clinical trial reporting standards. There are also linguistic means for ensuring coherence. These are often called markers of cohesion, following Halliday and Hasan‘s seminal work. Markers of cohesion can be on the level of words (lexical) or syntactic structures (grammatical). Examples from the blog entry so far include the parallel structure “good writing” … “good text”, anaphoric constructions such as “these”, “all this research”, and conjunctions such as “although”. Try removing them from the text and see what happens. How does it affect the flow? Does it become easier or harder to spot the connections between statements?
And I haven’t even gotten started about syntactic complexity, garden path sentences, and exotic vocabulary yet …
Armed with what I have told you so far about the active and the passive voice, about cohesion and coherence, and about the kinds of people who know about language comprehension, I would like to leave you with a little exercise:
Fisk the paper referenced in this blog entry by golden lady Pascale Lane. If you find it hard to get started, look at the references and count the number of references to the linguistic literature on texts, then divide them by the number of references to pontificating prescriptive position papers. You could also go to the literature review, and count the number of statements substantiated by actual experimental results. Feel free to report back in the comments.
For extra points, explain why the authors’ recommendation of standard ratios of active to passive voice in medical journal articles is poppycock and why any journal editors who follow it should be made to read William Topaz McGonagall’s complete works in one sitting, no toilet breaks allowed.
December 13, 2010
Language is often ambiguous. This ambiguity not only covers meaning, but also aspects such as emotion, attitudes, and speech acts. For the sake of this particular argument, let’s assume a situation where a reader R reads a text that a writer W has written. How the ambiguities in a text are resolved depends a lot on a reader’s world knowledge, past interactions between reader and writer, and the context in which the piece has been published, among other factors. Another complicating factor is the absence of intonation, the melody and rhythm of speech, phrasing and pausing.
When analysing a piece of text for ambiguities, we need to ask:
- What possible readings of a sentence exist?
- Are those readings resolved in the text?
- If yes, where – in the next section? In the next paragraph? In the next sentence?
When it comes to documents that are relevant to public health, such as information about treatments for cancer patients published by the foremost UK cancer charity, Cancer Research UK, we need to add three additional questions:
- Is it possible to get a reading from a sentence that sends an undesirable message?
- What sorts of readers are vulnerable to this undesirable message?
- Is the message so undesirable that the text needs to be revised to remove the ambiguity?
In a heated discussion on Twitter, some people, like Alice Bell and Rebekah Higgitt, have argued that the CRUK guidance on homoeopathy is acceptable – it points out the caveats, while ensuring it does not alienate people who use or are sympathetic towards homoeopathy. Others, such as Ben Goldacre, Dorothy Bishop and David Colquhoun have argued that the text sends some extremely undesirable messages and needs to be revised.
In the remainder of this entry, I will explain how some of these messages can be derived. Throughout, I will assume that the authors of the original web page wanted to convey the message that homoeopathy is not dangerous, and that homoeopathic consultations can even be helpful, but only when practiced by a qualified, reputable, conscientious therapist who sees their role as supporting, not supplanting, traditional treatment.
Before I go on to explain how these undesirable readings can be derived, it’s worth taking a look at the discourse of cancer in complementary and alternative medicine (CAM). Because cancer is such a frustrating and horrible and maddening disease, because treatment often seems worse than the disease itself, and because it is easy to feel lost in a complex system of machines and procedures, people with cancer are a prime target for CAM practitioners. Many practitioners are themselves cancer survivors who mistakenly credit all or a large part of their recovery to the CAM modalities they used during their treatment. Orac explains why this is a very easy mistake to make. (By the way, when it comes to cancer and CAM, it is hard to top Respectful Insolence, a blog I cannot recommend enough.)
Often, CAM practitioners can give people the kind of ongoing, empathetic support that can be difficult to get while in the medical system. However, there is potential for great harm if practitioners give advice that runs counter to that of the treating clinicians, or, even worse, if they actively lobby their clients to give up their conventional treatment. Sadly, such practitioners are not a negligible minority. Satisfied CAM customers such as the American Kim Tinkham, who may be about to die of her cancer, go on massive multipliers like Oprah to sing the praises of their particular “alternative”, dissuading countless others from life-saving treatment. In her book Smile or Die, Barbara Ehrenreich illustrates the intricate synergies between CAM therapies and the can-do, self-help, positive-at-all-costs attitude of many people with cancer in the US. Think yourself healthy! If you die, it’s your fault.
So what, Maria. These are all Americans. We’re British, we’re more sensible than this. Besides, these practitioners you speak of are just a few bad eggs who should be slapped down hard by their regulators. The web page quite clearly states:
It is very important that you have homeopathic treatment from a qualified therapist.
(Back to top)
Finding a homeopath
If your doctor refers you to one of the homeopathic hospitals you won’t need to find a homeopath privately. But if you are looking for private treatment, you need to make sure that you have chosen a qualified and reputable homeopath.
Currently in the UK, there is no single professional organisation that regulates homeopathic therapists. Therapists can join several associations. There is no law to say that they have to, but most reputable homeopaths do belong to one of the useful organisations. The best way to find a reliable therapist is to
Contact one of the organisations below and ask for a list of therapists in your area
- Ask the therapist how many years of training they’ve had and how long they’ve been practising
- Ask them if they’ve treated cancer patients before
- Ask if they have indemnity insurance (in case of negligence)
For more information about finding a reliable therapist and the questions you should ask look in our about complementary therapies section.
But it leaves out the most important question: Ask them whether they would ever knowingly give you advice that overrules that of your treating clinicians . (Also, run, don’t walk, if they start talking about their success rates.) For paediatric homoeopathy, the equivalent question would be “Do you support childhood vaccinations?”
There is also a problem with the wording of the second item of the list, because it is not clear what practitioners can claim to be able to treat. “Treating cancer patients” can mean “treating side-effects of cancer such as low mood” or “treating cancer itself”. The preceding paragraphs favour the first reading, but nowhere in the text is the second reading explicitly excluded. Readers are expected to make the correct inference themselves.
Personally, I would be very wary of any CAM practitioner who claimed they could “treat” cancer, and if a practitioner uses that expression in conversation with a new client who has cancer, I would again strongly advise the prospective client to run. All they can do is work with people with cancer or support the conventional treatment of people with cancer using complementary methods, and this is the wording I would recommend. Since homoeopathy is so badly regulated, it is all the more important to give people who insist on using it the right tools to find somebody they can usefully work with – so why is the ambiguity left in? Why are the all-important caveats never made very clear and explicit in a text that is almost certainly written to be accessible to people with a low(ish) reading level?
Next, let’s have a look at this gem:
More than 100 published clinical trials have looked at how well homeopathy works in treating various illnesses and symptoms. None of these trials give any scientific evidence to prove that homeopathy can cure or prevent any type of disease, including cancer. Many people say that homeopathy has helped their symptoms and some small trials have shown that homeopathy can have a positive effect. But some doctors and researchers have concerns about the way in which the trials were carried out. Below is a summary of recent research into homeopathy for cancer.
So, there’s no evidence that homoeopathy can cure or prevent, but trials have shown that there may be positive effects. Now stop and read the second-to-last sentence aloud. What readings do you get when you put particular stress on “some”? Which ones do you get when you stress “concerns”? Is “some” correct, or shouldn’t it be “doctors and researchers who are not homoeopaths themselves”, complete with implied “all”, or, as a compromise, “most”. After all, not so long ago, there was a certain report by a parliamentary committee which definitely seems to point in that direction. (Also, how many is “many people”, and in reference to which set is the quantifier defined?). It may be argued that we’re splitting hairs now, and that this ambiguity is not really relevant.
Well, what about the final piece, which again I will quote in full context:
Homeopathy is one of the most common complementary therapies used by people with cancer. As with most types of complementary therapy, people use it because it makes them feel better or more in control of their situation. Some people choose homeopathy because it offers a completely different type of treatment compared to conventional medicine.
Homeopathy for people with cancer is promoted as a natural way to help you relax and cope with stress, anxiety, depression and control other symptoms and side effects such as pain, sickness and tiredness. Homeopaths believe it can treat a wide range of symptoms and conditions.
The text does not state explicitly that homoeopathy is also promoted as an alternative treatment for cancer by unscrupulous practitioners, and that such practitioners are unethical. This is implied and can be easily derived from the web page if you read it in a sufficiently sceptical frame of mind – it does state in the following paragraph that there is no scientific evidence that homoeopathy can cure cancer. But another possible reading is that there is no scientific evidence yet, and if homoeopathy may well have effects, then maybe what the homoeopath says is true, conventional researchers just aren’t looking hard enough or in the right place; maybe this case – my case – will be the one that convinces the medical establishment that homoeopathy can indeed cure cancer? After all, doctors have tried to use homoeopathy to treat cancer! (The money quote: “In 2006, a review of 6 trials of homeopathy in cancer care could not find evidence that they worked in treating cancer.”)
I am not saying that this second reading is the intended, or indeed the default reading. What I am questioning is whether it should be possible to get this reading at all, and why there is no clear statement anywhere on the page that it is unethical for a homoeopath to claim to be able to treat cancer, and that making such a claim is a massive red flag.
The generic web page on CAM, which is two clicks away from the homoeopathy web page, is similarly open to unfortunate readings. On this page, it says:
A complementary therapy means you can use it alongside your conventional medical treatment. It may help you feel better and cope better with your cancer and treatment.
An alternative therapy is generally used instead of conventional medical treatment. All conventional cancer treatments have to go through rigorous testing by law in order to prove that they work. Most alternative therapies have not been through such testing. Some are not completely safe and can cause serious side effects.
What these paragraphs omit is that any complementary therapy can become “alternative” in the hands of the wrong practitioner, and that quite a few complementary therapists will advocate alternative therapies, even though they may not practice them themselves. In this way, complementary therapists can be the “gate keeper” or “entry drug” for the really serious alternative approaches. This fact is partly acknowledged in the section that states
A good complementary therapist won’t claim that the therapy will cure your cancer. They would always encourage you to discuss any therapies with your cancer doctor or GP. Complementary therapies are available from many different types of people and organisations. Later in this section you can find out where you can have complementary therapies.
But in the following section on alternative therapies, all this good work is obliterated with a single stroke when they state “Unlike complementary therapies, alternative therapies are used instead of conventional medical treatment.” The distinction between the two is not categorical, it is gradual, and the line between good and bad depends not on the treatment, but on the qualifications and ethics of the practitioner.
Good nutrition is important, curing cancer purely through nutrition is dangerous nonsense, but there are many steps inbetween, such as wheat-free dairy-free meat-free diets that may be recommended unthinkingly, leading to severe nutritional deficiencies.
A positive attitude may help, but thinking your tumors away is dangerous. In fact, the particular type of CAM described in the book by Brandon Bays I just linked to, “The Journey”, is a great example of the slippery slope. The Journey is a lovely guided body-centric meditation with elements of counselling. In itself, it is quite harmless. The danger is the ideology behind it, which the practitioner may choose to downplay. The Journey was how Bays healed herself after developing a massive tumour. It’s not clear whether the tumour was indeed cancer, but what is abundantly clear is how important it was for Bays to “clear” this herself, and how passionately she resisted the attempts of her doctors to treat her. From there, it’s just a short jump to seeing Kim Tinkham on Oprah, reading “The Secret”, and feeling you let the sisterhood down if you let those evil doctors poison you with their chemotherapy.
While we’ve been discussing lots of different readings up until now, all of this may feel very detached and theoretical – especially if you belong to the group of people who get the (possibly intended) reading without any problems. It is possible to devise a reading comprehension study that tests what sort of statements people think are supported by this text; all of the readings I have advanced can be reformulated into testable hypotheses. A carefully run study along these lines would surely be useful, involving people both with and without cancer, and with and without previous exposure to CAM.
Ultimately, whether any of these ambiguities or multiple readings are sufficient cause for concern to put public pressure on Cancer Research UK to change the wording is not for me to decide – I’m not a doctor, nor am I a medical researcher. I’m just a linguist. And I’m worried.
December 12, 2010
I am busy writing a grant, while not procrastinating on Twitter, that will fund a large number of well-funded PhD and early postgraduate positions. It is important to both me and the prospective funders to show that we provide adequate support for Fellows from underrepresented groups, such as ethnic minorities, people with disabilities, or female researchers. (I know, it sounds like Equal Opportunities bingo, but I’m trying to be both realistic and comprehensive.)
Given the recent debate on tuition fees, I might expand this definition to include other groups (working class, non-academic family, etc.) – any suggestions for a good phrase to use or further groups to include?
One of the ways in which we hope to help people from underrepresented groups is social media. From what I have seen, blogs / twitter / facebook and related networks are a great way for people to connect and support each other. Are there any formal studies (qualitative or quantitative) that I could cite to support the argument, and any good summaries of strategies that have been shown to increase the percentage of people from underrepresented groups that will stay on after their first degree?