Blending in English and Latvian compared

Blending is attested across a variety of typologically diverse languages. This word-formation process exists in both English and Latvian, however, its frequency and productivity differ significantly. Blending in Latvian is relatively rare, primarily due to the inflectional nature of Latvian. By contrast, clipping, is a common word-formation strategy in both languages, often in conjunction with compounding or the use of semi-affixes. 

A precise definition of blending remains contested among linguists. Some classify a word as a blend if it includes a single clipped component, whereas others argue that phonological telescoping is essential. A prototypical approach is often used to navigate these conflicts (Renner, 2012: 3), but a more rigorous delineation of blending remains needed (Bauer, 2012: 19). 

In Latvian, the creative and productive use of blending began to emerge towards the late 20th century, largely due to English influence, suggesting a word-formation pattern borrowing (Sakel, 2007: 15). This English-induced expansion parallels developments in languages such as Polish (Konieczna, 2012) and Slovene (Sicherl, 2018). Until the 21st century, most borrowed blends in Latvian were transparent or semi-transparent internationalisms, retaining recognizable structures and meanings, as in stagflācija (stagflation). 

Several types of blending can be identified in both English and Latvian. True blends include English guestimate (guess estimate) and Latvian okupeklis (okupācija piemineklis - occupation monument). Back-clipping compounds include English sitcom (situation  comedy) and Latvian nacbols (nacionālaisboļševiks - national Bolshevik). Front-clipping compounds are rare in English, as in podcast (iPod broadcast), and absent in Latvian, where podkāsts was adopted as a loanword. Other patterns include compounds with a back-clipping and a full word, such as English eurofighter (European fighter) and Latvian santehnika (sanitārā tehnika - sanitary technology), a structure widely used in Latvian. Compounds with a full word plus frontclipping: English motorcade (motor cavalcade), Latvian kafijholiķis (kafija alkoholiķis - coffee alcoholic) are not as frequent in Latvian and tend to have a few standard splinters. Both languages use splinters that function as semi-affixes (Lehrer, 2006: 592), including elements like -cracy, -holic, -gate, eco-, and euro- . An interesting distinction appears with semantically similar elements that differ structurally; for example, phobia functions as a full word, whereas -phobe is a suffix; as a result homophobia and homophobe fall into different linguistic categories.

The frequency of nonce blends varies, ranging from one-time instances to occasional widespread use, though the latter challenges the concept of a nonce blend. During the COVID-19 pandemic, a proliferation of nonce blends, particularly partial blends, appeared in both English and Latvian, incorporating elements such as covid-, -demic, and corona- (Rožukalne & Liepa, 2022).

The study allows us to conclude that blending mechanisms exhibit significant regularity and a high degree of cross-linguistic similarity. We think that blending should be understood in a narrower sense—primarily as telescoping—while still encompassing overlapping cases. These overlapping blends appear to be more predictable and recognizable, even without contextual support. Terminological fuzziness is further complicated in instances where blending borders on paronymic substitution of letters or sounds. Classical roots and international borrowings also play a prominent role in blend formation and enhance their spread. A distinctive feature of Latvian is that blending occurs exclusively with nouns. This is likely due to the inflectional complexity of Latvian verbs.

Corpus data reveal diverse but often unreliable frequencies for blends in both languages (NOW corpus and Latvian Tezaurs) While some blends are well represented, others are scarce or absent in corpora, although several infrequent examples have been documented by the author in both languages. This can to some extend be explained by the specific domains of blend use, e.g. advertisements which tend to be scarcely reflected in corpora.