Alex Steer

Better communication through data / about / archive

Everyone needs a really boring hobby

290 words | ~1 min

I've spent the last few days, in spare moments, looking at morphosyntax: how information about things like number, tense, person, and case are encoded in words, and how they should be processed and understood.

Which is obviously almost as rock and roll as throwing a TV out of a window. Even within the domain of linguistics it's pretty geeky stuff, I know. But it has the virtue of being interesting (to me) and pretty hard, meaning that understanding it usually takes a while and a bit of head-scratching.

And that's good. It's good for people who work in strategy, of any stripe, to have interests that make them think in pure, theoretical, or just plain logical ways. Strategists usually describe what they do as critical or rigorous or even logical, and they tend to use words like 'problem-solving' and 'analysis' pretty freely. But in a lot of what we do the real difficulty seems to lie in communication, or application, or the political difficulties of getting agreement on ideas and approaches. Much of the head-scratching comes from things other than really rigorous logical problem-solving: getting our facts straight and our numbers right. Which is not to devalue everything else - a strategist who can't get a strategy put into action is pointless - but just to say that the pure problem-solving part of the job can get neglected.

It's not just brain-training, either. Whether it's crosswords, number puzzles, computer games, rock-climbing or one-man tiddlywinks (not a euphemism), it's good to compete only with your own brain from time to time. If I don't understand some theory on language, that's no one's fault but mine; the satisfaction when something clicks is uncomplicated; and the challenge to understand it better is irresistible.

# Alex Steer (16/05/2010)


PowerPoint-bashing

388 words | ~2 min

Is PowerPoint making us stupid? No.

A piece in the New York Times (now a few days old) describes the agony of US soldiers who spend too much of their time (a) making or (b) sitting through turgid PowerPoint presentations. The piece is good and the strategy diagram at the top is funnier for not being a joke. But there's a nagging suggestion, despite the disclaimer that 'no one is suggesting that Powerpoint is to blame', that PowerPoint is to blame - not just for military failings but for much that is wrong within the domain of strategic planning. This is pretty typical:

Commanders say that behind all the PowerPoint jokes are serious concerns that the program stifles discussion, critical thinking and thoughtful decision-making. Not least, it ties up junior officers - referred to as PowerPoint Rangers -in the daily preparation of slides, be it for a Joint Staff meeting in Washington or for a platoon leader's pre-mission combat briefing in a remote pocket of Afghanistan.

The ubiquity of PowerPoint from boardroom to war room is not new, nor are the mutters of concerns about this. Edward Tufte's essay 'The Cognitive Style of PowerPoint' examines some flaws in how PowerPoint organises information by default (essentially paratactically - 'one damn slide after another' - with rigid nested hierarchies by means of bullet points).

It's quite a leap to assert that PowerPoint is killing the art of strategic thinking, though - by which I mean thinking that is itself well fitted to the problems it is trying to solve. I'd venture that rigid organisational hierarchies, not hierarchies of bullet points, are more at fault, as is simple time pressure. Strategies take time and space, and the crystal clear thinking that results can come out of some surprisingly unordered processes (ones strategists tend not to admit to, which is a shame). Bad PowerPoint is a symptom of stifled strategic thinking, not a cause. Sure, the software doesn't go out of its way to offer ingenious ways of concatenating points of information, but that's well beyond its remit. If the paper-clippy Office Assistant starts thinking up my strategy for me I'll know it's time to lay off the coffee.

# Alex Steer (16/05/2010)


Using linguistics to debate ethics

177 words | ~1 min

It's common, when teaching ethics and policy issues, to take examples from the sciences, especially biology. You probably know the type of question used to spark these debates. Should human cloning be legal? How about GM food production? How much regulation is appropriate for a DNA database?

These questions are relevant and interesting, and provide a good applied basis for developing one's theoretical ethical and political analysis skills. Oddly, though, I've never seen linguistics used as the jumping-off point for such exercises, despite the obvious applicability of case studies like the following:

A university has 25,000 students and provides each of them with an email address, and the university's mailservers store copies of all mail sent and received by students. Should it be legal to make the text of those email messages accessible to:

  1. The university's own researchers?
  2. Researchers from other universities?
  3. Market research companies?

In each case, should the data be anonymised?

Or:

Should sentiment analysis software be used to target advertising to individual users of Twitter or Facebook?

# Alex Steer (16/05/2010)


A few from the last few weeks

35 words | ~0 min

There's been a bit of a hiatus on posting here, as I've been busy. I'll shortly be adding several posts that I've written, but not uploaded, over the last few weeks. Sorry for the bombardment.

# Alex Steer (16/05/2010)


Search and economy

352 words | ~2 min

Just reading a case study by the NLP (Natural Language Processing) team at PARC on improving search string parsing in Powerset. It's good, but it includes the following very common question:

CEO Barney Pell observes... 'Why do we have to translate our intelligence into a grunting pidgin language in order to interact with computers?' To address this gap, Powerset's founders decided to create a consumer search engine based on natural language processing technologies, which enable people to interact more naturally with computers through normal language expressions instead of forced computer jargon.

This classic defence of NLP for search (which scarcely needs defending, as its benefits are so obvious) is based on the assumption that people want to type natural language expressions into search engines in full. This is based on the correct insight that severe restrictions on string formation (think Boolean operators or Google's inurl: or filetype: operators) can be offputting, but it's an overextension of this insight which ends up assuming that people's time has no value. Compared to computer scientists, who will do almost anything to save on an unnecessary keystroke, general searchers are pretty forgiving. Still, it would be unwise, surely, to pour too many NLP resources into high-end sentence processing of the kind needed to work out the information demand in a string like 'Which is the most expensive restaurant in New York where a fish course is not served?', if most search strings are not formed like that. Better to spend time, money and effort understanding how searchers economise - in other words, to treat search strings as a corpus and analyse the features which over-index in that corpus compared to a corpus of general written English. Click-through rates could be used as a register of semantic intent.

Of course, higher-level sentence processing is useful for answering the queries of users who do search using full sentences, but search strings probably count as a sociolect in themselves whose grammar is worth understanding.

# Alex Steer (16/05/2010)


Electioneering up the garden path

917 words | ~5 min

It must, in fairness, be a nightmare writing party election manifestos, knowing that every word will be pored over by journalists, lobbyists, organisations and individuals looking to wring every drop of significance from whatever you commit to paper. Little wonder they're frequently so dry - any misplaced phrase can drop with a resounding clang.

In the run-up to the UK general election, though, the Labour Party has surpassed itself with the innocuous but amusing title of its manifesto. The Conservatives' manifesto (PDF), launched today, dodges most ambiguities and seems decisive by using imperative phrases for most of its section headings - e.g. 'Make politics more accountable', 'Combat climate change' (though it falls at the last hurdle with a section on 'One World Conservatism' which, for me, activates a sense of similarity not with the happy notion of a 'One World' movement but the sinister prospect of 'One World Government').

Labour's manifesto, though, sticks its ambiguity on the front page.

Labour manifesto - A Future Fair For All

Fans of psycholinguistics will be familiar with the garden path model of sentence processing, which holds that readers parse lexical units (words or compounds) in sentences sequentially, checking for semantic consistency as we go and revising our initial assumptions whenever we find something that confounds them - so we can read and make sense of the phrase Jack gave in and parse 'gave in' as an intransitive phrasal verb in the sense 'surrendered', then revise our reading when our eyes take in the next few words and we read that Jack gave in his homework, and realise that 'gave in' is a transitive verb in the sense 'submitted'. We might then see the word book at the end of the sentence, and then we'd have to re-parse 'homework' as an attributive noun (or an adjective, depending on your point of view) in the compound 'homework book'.

So what's the issue with A future fair for all? One adjunct to the garden path model is the idea of minimal attachment, which proposes that as readers or listeners we assume simple/familiar relationships between words until the evidence forces us to reconsider. So, for example, we find sentences like the horse raced past the barn fell hard to make sense of, and I put the candy on the table in my mouth faintly confusing, because we assume we are reading a sentence about a horse racing past a barn (rather than being raced past a barn and falling) or candy being put on a table (rather than candy already on a table being put in one's mouth).

Minimal attachment errors can also lead us to some odd mental images - such as a table in one's mouth. The community at Language Log christened these baffling sentences 'crash blossoms', and the Crash Blossoms Blog records plenty of examples, mainly from the linguistic pressure-cookers that are newspaper headlines.

A future fair for all seems to be causing its readers some trouble and amusement. While nobody seriously assumes that Gordon Brown is planning a massive funfair at a to-be-specified later date, it's clear that the phrase is being activated in this way for plenty of people, perhaps before its intended meaning - 'a future that is fair for everybody'.

Minimal attachment is to blame. Here's an example of the way a garden-path reader might parse the line.

A future (determiner/article + noun)

A future fair (determiner/article + adjective + noun: future reanalysed)

A future fair for all (determiner/article + adjective + noun + prepositional phrase)

(Note: it's possible to parse a future fair as determiner + noun + adjective, with 'fair' as a postpositive adjective - as in the old hymn which promises believers 'a harp, a home, a mansion fair'. Postpositive adjectives are now fairly uncommon and archaic, though, so this meaning is less likely to be activated.)

It takes another look and a bit of contextual thinking to realise that Labour are not promising a funfair in the future, and that the minimally attached reading is not plausible, even though it is semantically well-formed.

We've been stitched up by a feature of English syntax: the reduced relative clause. This feature allows relative clauses to skip the introductory relative pronoun - so instead of the shoes which are on my feet we can say the shoes on my feet.

This means that we can, without seeming to violate any constraints of English syntax, talk about a future fair for all. It's just that garden-path processing and minimal attachment will mean that what we intend may not be the first meaning that occurs to our readers.

Which, when it comes to canvassing for votes, is not a good thing.

# Alex Steer (13/04/2010)


Nestlé, Greenpeace, trends and impact

218 words | ~1 min

By way of a quick follow-up to the Nestlé Facebook social media, I thought it might be worth seeing the impact of the whole affair on the Nestlé and Greenpeace brands, and on the issue of palm oil in KitKats. This is pretty basic stuff - I'm using Twitter as my search medium, just recording mentions. If anyone has the kit to do some up-to-date sentiment analysis, it would be great to see (though I can guess). This from Trendistic:

Image lost in database transfer - sorry!

So, perhaps because of the ensuing Facebook fan page disaster, the affair seems to have driven talk about Nestlé more than it has talk about Greenpeace, but there's still been a credible showing for Greenpeace, who broke the story and have drummed up a lot of the publicity. (Interestingly, mentions of Nestlé spiked a day before mentions of Greenpeace, suggesting that this was first being identified as a story about Nestlé, not a story from Greenpeace, which is understandable.)

KitKats are also being mentioned a lot, over a similar timeframe to Greenpeace, suggesting the campaign message was being picked up on. The specific issue - palm oil - rates a lot lower though, so any follow-up from Greenpeace might want to reinforce that.

# Alex Steer (19/03/2010)


Nestlé, Facebook and tone of voice

529 words | ~3 min

Less than 24 hours since it happened, there's been plenty of coverage of Nestlé's rather fatal decision to tell contributors to its Facebook fan page that their posts will be deleted if they use profile pics that are alterations of Nestlé logos or product. (The post was prompted by protests on the fan page about the use of palm oil in KitKat bars. See here for useful context.)

The post has prompted an absolute storm of hate messages on the page, and rightly so. Nestlé is, frankly, not a brand that enjoys an enormous amount of consumer warmth. Fairly or not, it is still haunted by the memory of the baby milk issue, for a start. So being all heavy-handed would never get a good response. Brands using social media have no more clout, nor any more right to lay down the law, than any other account-holder. If they want to be part of the conversation, they have to be prepared for that conversation to be pretty frank sometimes, and they can't assume they can control it. This is definitely not rocket science. What I'm saying is so well understood by marketers that it is tiresome to have to write it, and Nestlé should definitely have known better.

But this is not just a PR blunder. It's very damaging for Nestlé's brand because it shows that they are not in control of their tone of voice online. The following exchange (pinched with thanks from Digital Inspiration) shows that whoever was moderating Nestlé's Facebook page at the time felt able to get away with a tone of voice that was sarcastic, hectoring and aggressive:

Nestle Greenpeace Facebook social management

Tone of voice really matters, and so does consistency. Some brands do it really well and end up sounding really distinctive within their categories - think of Ben and Jerry's, for example, or the new campaigns from Tesco Mobile or Kotex. But even if you can't craft an ingenious tone, the least you can aim for is an engaging, friendly and inclusive tone that you carry across all your communications. Nestlé's website is a model of fairly bland but clear and inoffensive tone. Its Facebook page has destroyed a lot of that tonal equity in less than a day.

This tonal collapse will be costly. For now, it demonstrates a lack of foresight, and a fair degree of contempt not just for the contributors to its Facebook page (who are now showing contempt right back) but for the medium itself, implying that the brand can't be bothered aligning its tone with the rest of its communications output, and by extension that social media and the people that use it don't much matter.

Disappointing.

# Alex Steer (19/03/2010)


Conversation analysis by iPhone

367 words | ~2 min

There's a branch of linguistics called conversation analysis (CA) which, as its name suggests, analyses the dynamics of conversation. Whereas many areas of language study (including discourse analysis, of which CA forms a part) focus on language structures within individual utterances or texts, CA is useful for giving us insights into how language use defines and reflects the relationships between speakers. So CA focuses on aspects of speech like turn-taking, interruption and repair (e.g. repeating or clarifying when something is unclear), and making sense of their use as interpersonal tactics. For example, if I'm constantly interrupting you and not letting you get a word in, that's a pretty aggressive strategy to assert my status. If I'm constantly deferring to you, I'm acknowledging your status. If you say something stupid and I just sit there and look at you, well, you know you've said something stupid. And so on.

So it's good to see there's now an iPhone app which can do some basic CA in real time. Talk-o-Meter can distinguish between voices in two-way conversations and can track who's hogging the chat.

Talk-o-meter app

Pretty fun, even if it's likely to make you so ridiculously paranoid about how much you're talking that you'll completely skew the results and both end up deferring to each other like the vultures in The Jungle Book.

On a side note, if 'markets are conversations' as the Cluetrain Manifesto and everyone who's ever read it insists on pointing out, then a surface knowledge of conversation analysis might be a useful thing for marketers to have, especially if they're helping brands use social media effectively. Take a look at the best brands on Twitter or Facebook and you can see some of the basics of polite conversation at work - turn-taking, face-saving, deferral, good humour. Take a look at the worst, and all you see is a one-way stream of self-promotion that, in any conversation between people, would lead to you spitting out bits of iPhone for the rest of the week.

# Alex Steer (19/03/2010)


St Patrick and the prescriptivists

230 words | ~1 min

17 March was St Patrick's Day, and this obviously led to an upswing of online chatter about the fifth-century Romano-British snake-chaser.

On Twitter, Marcus Campbell (@scronide) noticed, and remarked, that a lot of US tweeters were referring to St Patrick as 'St Patty', in contrast to 'Paddy', the commoner abbreviation in British and Irish English. In response, he set up paddynotpatty.com, a light-hearted site campaigning against the use of the 'Patty' abbreviation. As the site announces in big white type:

It's Paddy, not Patty. Ever.

But the internet disagrees. Look at Twitter:

Images lost in database transfer - sorry!

The short period in which 'St Paddy' outscores 'St Patty' may reflect the time difference between the British Isles and North America.

Martin McGrath (@martinmcgraththought it was a new phenomenon. It's not, as Google Trends shows:

Image lost in database transfer - sorry!

Which is a shame, as the Paddy Not Patty website's hilarious. But the commitment to record language as it's used never takes a day off, not even to paint its face green, drink Guinness and have a bit of a fight.

# Alex Steer (18/03/2010)