The latest opinion polls seem to indicate that the Irish referendum on the Treaty of Lisbon will be very close.
For some reason, most referendums tend to end up with a close result, even if one of the sides is far ahead before the campaigns start.
I’ve been thinking a bit about this, and I think an explanation might be found in the Grice’s conversational maxims.
Basically, the maxims are some rules that form the basic assumptions underlying human communication. The rules are often broken, but not without reason, and breaking them implies a meaning on its own.
In particular, I think the Maxim of Relation (“Be relevant”) is crucial here. In means that one will assume that a question asked is relevant, that is, it is assumed it can be there is more than one possible answer. If the politicians arrange a referendum but say that only voting yes really works and that a no would be a disaster, they break this maxim, which makes people confused and angry and they start thinking they’re being lied to. They might even think the Maxim of Quality (“Be truthful”) is violated, too.
To make it even worse, people tend to say no when they don’t fully understand a question, which is why making people read the whole treaty is normally a winning strategy for the opponents.
Because of all this, I believe referendums only work where politicians are happy to proceed with either outcome. For instance, Scotland can feasibly continue as a part of the UK or become an independent country, so this is a possible topic for a referendum.
On the other hand, saying yes to a complex treaty that is a messy compromise between 27 countries that took years to negotiate is clearly not a good basis for a referendum. If one had wanted one, it should have been held years ago and have been about the negotiation mandate for the Irish government (but that’s of course not what the Irish constitution demands).