The Wall Street Journal ran an article the other day about a new profanity policy. The policy is a spinout from the public embarrassment they got during a Senate hearing back in April. Some of the traders’ blunt and explicit comments about the securities they were selling were read on the Senate floor. (It was the first time I’d heard CSPAN bleeped out.) Ignoring the ethical issues of selling a product that you don’t believe in, Goldman is trying to reduce the potential for future embarrassment by cleaning up their language before the next time.

I have mixed opinions about the new policy. On the one hand, that industry has a very macho image. Profanity is an ingrained part of their culture. Profanity recognizes and reinforces the aggressive attitudes valued among the traders. Profanity can show the passion of the speaker. And, arguably, it helps in bonding and cultural norming. Similar trends are common among soldiers, journalists, police, some sports teams, etc. The language is offensive to outsiders but, in some ways, that’s the point. It becomes part of the group identity. And as long as it’s limited to the insiders who participate by choice, well, you should be cautious about changing the a successful culture.

Having said all that, I think the new policy is a good one. Clearly their behavior has gone too far. It was adversely impacting the business and needed to be reined in. More than that, the informal language leaked out of mere speech and into their emails, creating a permanent record that will inevitably be exposed to outsiders who do not participate in, understand or appreciate the ingroup’s culture – outsiders who may be deeply offended by the choice of language. That’s just inexcusable.

As we’ve often talked about before, emails are official business communications and must be treated as such. They deserve all the thought and professionalism that we used to put into a formal memo back in the days of carbon paper and typewriter ribbons. If you’d be embarrassed to have your email read in church or quoted on the front page of the newspaper, then you should rethink the message.

But I’m not such a fan of the automated filters that Goldman and others are using to enforce their policy. Profanity filters try to identify the offensive words and, depending on the company’s settings, return the email to the sender, block the email or allow the message to go through but flag a copy to HR. The filters use long lists of keywords, usually including common abbreviations and aliases (like adding ** in place of the vowels). The problem is that the offensiveness of a message is often dependent on context. As soon as you get a list long enough to be even marginally effective, you will inevitably suffer false positives.

As an example, my company tried to do something similar as a spam filter a few years back. In hindsight, it’s not really a surprise that construction companies (many of whom were our customers) use the word “erection” in legitimate business messages. BS can be a pejorative abbreviation or a respectable undergraduate degree. POS can describe a defective piece of hardware or your Point Of Sale register (and, yes, your POS can be a POS if you bought from the lowest bidder).

I should note that some of the most advanced filters now claim to be able to differentiate meaning based on the context of the message. They do alright for spam filtering and are showing promise for some other purposes but I don’t think they’re ready for use as profanity policy enforcement. The English language is too loose and our people are too creative. Very few of the filters would correctly parse the paragraph above and none can keep up with the changing acronyms and innuendo that people employ to dodge the censors. My prediction is that the filter will have some short-term shock value but the real change will only come when managers do their jobs – teaching employees the new standards, leading by example and holding people accountable when they backslide. That’s the only real way to change the culture.

Leave a Reply