The following describes the general criteria which were used to assign a word to a particular Borrowed status. An effort was made to make sure that there was some consistency to the way these categories were employed, but there was necessarily also a degree of subjective assessment as well. The descriptions as to how each category was used will cover most words, but, of course, they all “leak” to some degree.

Clearly borrowed:

Generally, I classified a word as clearly borrowed if (i) it did not seem possible to propose a reasonable English etymology for the word and (ii) it was possible to propose a non-English etymology for the word which was a good match for the relevant Saramaccan word both formally and semantically. When a word did not meet these criteria but was still classified as clearly borrowed, a justification for this will generally have been given in the Comments on Borrowed field. The most common reason for such exceptions was when there was an imperfect, but still good, formal and semantic match of a Saramaccan word with a word of Afri-can origin, since available sources on the relevant African languages are limited and the lack of a better match could easily be due to gaps in documentation.

Probably borrowed:

Generally, a word was classified as probably borrowed when there was a good proposal for a non-English etymology for the word, but there were (i) important open questions about either the quality of the proposed etymology on formal or semantic grounds or (ii) there was a competing English etymology which seemed unlikely but which I could not rule out entirely. In most cases, there should be a comment in the Comments on Borrowed field briefly explaining the nature of the problems resulting in this classification.

Perhaps borrowed:

Some words given this classification, broadly speaking, were similar to words in the probably borrowed class except the nature of the relevant problems was more significant either because there was more than one distinct kind of notable problem (for example, issues with both the formal and semantic match for the proposed source word) or because one single problem raised particularly significant issues (for example, the semantic match may have been especially tenuous). Another class of words given this classification involved compounds (and other polymorphemic structures) composed of two elements from the same non-English source language (e.g., Gbe languages or Portuguese) where I had no evidence indicating whether the relevant complex structure was formed internally to Saramaccan or if it was borrowed as a unit. In most cases, there should be a comment in the Comments on Borrowed field briefly explaining the nature of the relevant problems.

Very little evidence for borrowing:

The majority of the words in this class are cases where there is no clear English or non-English etymology for the word. Since, the Saramaccan lexicon is assumed here to have branched off of the English lexicon, any word without a clear English etymology is a candidate as a loanword. However, the lack of an English etymology is weak evidence for borrowing, especially given that we can expect that Saramaccan would have formed some new words completely internally. For this class of words, there will be a comment in the Other Comments field explaining this briefly. Another class of words in this category were those that had a good English etymology but where another etymology (e.g., Dutch or Sranan) also seemed possible, but relatively unlikely. In most cases, there should be a comment in the Comments on Borrowed or Other Comments field briefly explaining this.

No evidence for borrowing:

As discussed in the Saramaccan chapter, for the purposes of this vocabulary, the Saramaccan lexicon is viewed as a highly divergent variant of the English lexicon. Accordingly, monomorphemic words with good English etymologies are considered to give no evidence for borrowing (except, of course, for words for relatively modern concepts). In addition, for the most part, morphologically analyzable words (including compounds) are generally considered to offer no evidence for borrowing unless there is specific evidence indicating they are not the result of word-formation processes within Saramaccan.

The ages were generally enumerated into descriptive categories of lexical strata to which specific years were assigned to at a later stage, with the year spans for different strata completely over-lapping in some cases. This provides a relatively coarse-grained view of the ages of the various lexical items in the database. Some words, undoubtedly, could be given more precise ages, but such work was not attempted here. Age spans were all rounded to fifty-year intervals so as not to give any mistaken impressions of precision for their beginning and end points.