• borrowing processes
• indirect borrowings
• Latin
• Naturalness Theory
• Old English
• system-dependent naturalness
• system-independent naturalness

The present work demonstrates that indirect borrowing processes have more profound effects on the borrowing language. Gusmani (1986: 287-288) notes that:
- semantic loans alter the structure of meaning;
- loan-formations establish relations between form and meaning which were non-existent before;
- the borrowing of morphemes influence the grammar of the language.
In the present research I have tried to prove that indirect borrowings are universally more natural than direct borrowings. From the point of view of system-independent naturalness, indirect borrowings are the most natural borrowings, and direct ones are the least natural. Among indirect borrowings, LTs are the most natural borrowings since they show a high degree of iconicity, morphotactic and morphosemantic transparency, biuniqueness and morphological and semantic indexicality. LRs are less natural than LTs because they are less iconic and indexical than LTs, even though they may be equally transparent and biunique. LCs are even less natural than LTs and LRs, since they are less iconic and have a lower degree of indexicality than LTs and LRs, even though they may be equally biunique and transparent. SLs (analogical and substitutive) are less natural than LFs and direct borrowings because, in the case of simple words, they totally lack morphosemantic and morphotactic transparency. But even though they are morphotactically and morphosemantically transparent (in the case of complex words), they are non-iconic and lack biuniqueness. As for indexicality, it is only semantic and is rather weak (or non-existent) in the case of substitutive semantic loans. Still, SLs win out over indirect borrowings on the principle of linguistic economy. Direct borrowings (nativised and non-nativised) are less natural than LFs since they are morphotactically and morphosemantically opaque, even though they may be equally biunique and iconic, and have a high degree of phonetic indexicality. They also lose against SLs on the principle of linguistic economy, as has been mentioned above.
The Old English data confirms the scale of system-independent naturalness of borrowings. The only exception are SLs, which seem to be viewed as more natural than LCs and direct borrowings in Old English, and consequently are more frequent. This is due to the fact that in Old English, as in many other inflectional languages, the naturalness on the parameters of transparency, iconicity and biuniqueness is often sacrificed for the greater naturalness on the parameter of linguistic economy.
Two transitional phenomena, hybrid loan-formations and phonologically-triggered loan-formations (or folk-etymologies) clearly show the general tendency of borrowings in the direction from a less natural direct borrowing towards a more natural indirect borrowing.
Of course, the Old English data is not enough to prove the validity of the system-independent scale of naturalness, proposed in this thesis. A cross-linguistic study in the field of borrowings is necessary. This may become one of the future perspectives for the study of borrowings. The present research has provided empirical support for predictions regarding Old English borrowings and an explanation of the preference for indirect borrowings in Old English as opposed to the preference for direct ones in Middle and Modern English.