This book traces a pesky idea that's been bumping around Western culture for centuries: the idea that a language (or language-like formalism) is possible (which either existed, or which we can devise) which is somehow truer than our mundane languages. Eco traces this idea starting from its roots in ancient times, and he goes into fine detail in discussing the "philosophical languages" of the Renaissance, before discussing more recent constructed languages (Esperanto and the like).
The prose is very clear and straightforward, and the subject full of interesting nooks and crannies.
The book is most valuable in that, once you've read it, you will start recognizing the "perfect language" idea popping up everywhere -- the idea that if we just stick to a really rigid formalism (which we're /almost/ finished coming up with!), then we can get everything right. This idea appears in everything from formalist linguistics ("since the framework is perfect, you just plug in the right parameters for your language, and it works!"), to the voodoo equations of quantitative political science ("and this formula /explains/ why the Sino-Japanese war happened!"), to American law ("I don't care if this law is just -- I'm talking about whether, formally, it's Constitutional; because that's what really matters!"), to the endless wars over which is the best programming language ("Python is better than Perl because it's based on objects, and if you don't understand why that's important, you need to learn more lambda calculus, and indent your code more /correctly/!").
It'll make you think twice about anything that needlessly uses a formalism for expressing what could be said just fine in one of these mundane languages we speak!