A recording medium, method, and apparatus for reproducing text subtitle streams are disclosed. A microcomputer initially reads a set of PlayItems and a SubPlayItem included in a PlayList which is recorded on a recording medium, and it further reads synchronization information included in the SubPlayItem. Each PlayItem is configured to reproduce at least one main AV stream, and the SubPlayItem is configured to reproduce a text subtitle stream. The synchronization information indicates a presentation start time of the SubPlayItem on a global time axis of the PlayList. When the PlayList is randomly accessed at a random access time on the global time axis, the microcomputer calculates a corresponding access time on a time axis of the SubPlayItem. Then a text subtitle decoder decodes for presentation a dialog presentation segment with the closest dialog start time after the corresponding access time on the axis of the SubPlayItem.