For us humans, language mediates the world. If so, why not caption it all? GPS linked to the WWW makes it feasible to tie unlimited text to specific places. Go to a cultural site, open the text referenced to that space, read it-- add your own annotations--make new links. Thus the space/object can be extended in time. (This was posted in February; Google has sonce introduced several aplications that bring it close, of not to, reality).
An entirely different way of approaching this is to reflect on urban life without text--no street signs, no text advertising, no "restroom", no license plates on the cars, no elevator button numbers, no bus, subway or emergency pull handle directions. Add to the list of banned items the quasi-texts of handicapped symbols, corporate logos and direction arrows. Maybe strolling through midtown Manhattan, you'd be able to decode the architecture in some places (Public Library), but for the mosrt part, urbanity is text.