Relativized it. ke'a is a pro-sumti (meaning it takes the place of a fully-specified sumti). ke'a is only used inside relative clauses, where it indicates the place in the relative clause's bridi that the speaker intends have apply to the sumti the relative clause is attached to. ke'a can generally be represented by the English word "it". In the case of nested relative clauses, ke'a or ke'a xi pa refers to the inner-most sumti to which the relative clause containing it is attached. ke'a xi re refers to the next outer-most relative clause attached sumti, and so on for higher numbers.
le stizu poi ke'a srana mi cu blanu
The chair which it pertains to me is blue.
My chair is blue.
Sumti question. ma is a pro-sumti (meaning it takes the place of a fully-specified sumti). ma is the sumti question marker, and represents, in many cases, any of the English words what, who, where, why, when, and how. Specifically, it turns the entire bridi into a question, and requests that the listener provide a sumti as an answer. The sumti response fills the place where ma was in the original bridi and the resulting bridi is the answer offered to the question. To indicate that there is no value that could make that bridi true, use no da. A more extreme rejection of the presuppositions or implicatures involved in a question is to reply with na'i. A bridi with more than one ma can be responded to with an unconnected string of sumti.
le verba vi ma pu cadzu le bisli
The child near what walked on the ice?
Where on the ice did the child walk?
Nonexistent argument place. zi'o is gramatically a pro-sumti, meaning it fills a sumti place, but unlike other pro-sumti zi'o actually removes the place it fills from the bridi it is in entirely. A bridi with zi'o in it actually represents a completely different relationship, one with one less element being related.
(The "maker" place of zbasu has been removed, so a maker is neither denied nor asserted.)
(Asserting that, regardless of how the pencil does its marking, the concept of pencil does not require there to be an eraser.)
The meaning of -zil- has had some confusion, and this ties in nicely to the current debate as to the meaning of ziltau.
A bit on what the CLL (doesn't say/)says regarding the matter.
zi'o with sumtcita
Unspecif it. zo'e is a pro-sumti (meaning it takes the place of a fully-specified sumti). It represents an elliptical or unspecified value. It has some value which is irrelevant or obvious in the current context. All empty places in Lojban are implicitely filled with zo'e, making it (by far) the most-used word in the language, in a sense. zo'e can represent just about anything. The important exceptions are no da, which is equivalent to putting na in front of the selbri of the bridi in question and hence alters the meaning completely, zi'o, which utterly changes the nature of the bridi to one which has a different place structure, and ma, which turns a statement into a question. zo'e can represent a referant of any complexity. To fully specify the thing represented by zo'e may require very complex Lojban, including abstractions, relative clauses, relative sumtcita, and combinations thereof.
Note that, technically, every single bridi without every place filled in is a usage of zo'e.
("there" is being used to gloss zo'e in this case.)
Typical it. zu'i is a pro-sumti (meaning it takes the place of a fully-specified sumti). zu'i represents some value that is typical for the bridi place it fills. It can represent sumti of any complexity, including abstractions, relative clauses, relative sumtcita, and combinations thereof, but whatever it represents must be typical for the bridi in question. If the bridi relationship does not actually hold for any typical thing in the place filled by zu'i, then the presence of zu'i in that place makes the bridi false. Lojban constructions that zu'i explicitely cannot match include no da, which is equivalent to putting na in front of the selbri of the bridi in question and hence alters the meaning completely, zi'o, which utterly changes the nature of the bridi to one which has a different place structure, and ma, which turns a statement into a question. None of these can be typical place fillers in any case.
We can't know what the typical route is without knowing (at least) the layout of the house, but the last zu'i probably can't be anything other than lo nu cadzu, which is "walking".
This sentence is actually a meta-linguistic statement about the nature of two Lojban gismu, srana and ckini, and as such loses essentially all meaning in translation.
Presumably the typical value is six (a cube).