What You Actually Need to Know About Translating English into Modern Greek
I spent three years working on localization projects for a software company before I realized how much of the standard advice out there is just guesswork dressed up as expertise. Most English-to-Greek translation goes poorly because people treat it like a word-for-word swap. It isn't. The languages operate on completely different grammatical bones, and the gap between them is wider than most beginners realize. The article continues below Greek has a defnite article system that doesn't exist in English in the same way. You use the definite article with abstract nouns, body parts, and concepts that would never take one in English. « » literally means "the hand of mine" — you can't say just "my hand" the way an English speaker might think the translation works. It's not a quirk. It's structural.
I once had a client who sent me a technical manual with about 4,000 words. Straightforward engineering documentation. Two days later she came back asking why I had used the subjunctive mood in eleven places where she expected the indicative. The original English text had modal verbs like "shall" and "should." In Greek, those don't map cleanly. I'd chosen the subjunctive because the context described a condition, not a statement of fact. My client had no grammar training and just wanted it to sound "right." We spent another six hours going through each instance. I lost track of how many times I explained that Greek doesn't have a direct equivalent to the English modal system. That was also the project where I learned to stop trusting machine translation output for anything beyond rough scaffolding. Google Translate and DeepL will get you about sixty percent there on technical text. The remaining forty percent is where errors compound — case agreement breaks down, aspect shifts get lost, and idiomatic phrases come out literally, which in Greek usually means coming out wrong. The verb system is another thing that trips people up constantly. English relies heavily on tense and aspect through auxiliary verbs and timing context. Greek encodes aspect directly into the verb morphology. The aorist and imperfect distinction doesn't exist in English at all. An English speaker might write "I opened the door" and not realize that in Greek, whether you use the aorist or the imperfect changes the meaning in a way that matters. Aorist presents it as a completed event. Imperfect presents it as ongoing or repeated. Get this wrong and your translation reads like it was written by someone who doesn't understand what happened in the original text.
Word order is more flexible in Greek but not infinitely so. You can move around subjects and objects more freely than in English because the case system marks grammatical relationships. But free flexibility doesn't mean random. I've seen translations that place adverbial phrases in positions that are grammatically legal but sound distinctly unnatural to a Greek reader. The rule of thumb is: put the new information toward the end of the clause. Greek tends to move from given to new, same as English, but the margin is wider. A well-placed adverbial in the middle of a sentence might still work in Greek, but it carries a different pragmatic weight. Here's something nobody mentions enough about register. English has a relatively flat register system compared to Greek. The formal-informal distinction in Greek — versus — isn't just politeness. It signals something about the relationship between speaker and listener that English simply doesn't encode. When translating a business document from English, you need to decide early whether the target audience expects formal address throughout or if informal is acceptable. Getting this inconsistent reads as unprofessional in Greek the way it wouldn't in English. I default to formal in business and technical contexts unless told otherwise. That's been my standard for a long time now. Punctuation in Greek has a few specifics that cause problems. The question mark ; and the semicolon · serve different functions than their English counterparts. The Greek semicolon is actually a question mark. The dot-on-top · acts as a semicolon or sometimes a comma depending on context. If you're translating a document that includes dialogue or Q&A sections, mixing these up will be immediately obvious to any Greek reader.
Get the Full Details
For workflow, here's what I actually use. I start with a CAT tool — I prefer OmegaT because it's free and handles Greek well. I run the source through DeepL first as a draft, not as final output. Then I go through it clause by clause, checking agreement, aspect, and register. A typical 2,000-word technical document takes me about two to three hours from raw MT output to finished translation. The MT step saves maybe an hour and a half, but the real time cost is always the quality pass. No shortcut around that. One edge case that comes up more often than you'd think: proper nouns and brand names. Some brand names get transliterated, some stay in English, and some develop their own established Greek form over time. Amazon stays Amazon in Greek text. Coca-Cola is -. There's no consistent rule. I keep a running glossary of established forms and check against existing Greek-language corporate materials whenever I'm unsure. Making something up on the fly is the fastest way to create inconsistency across a project. Numbers and dates also deserve attention. Greek uses the decimal comma and the thousands separator is a space or non-breaking space, not a comma. Date format is day-month-year, so 15/3/2024 not 3/15/2024. I've caught these errors in peer-reviewed work, so don't assume they only happen in amateur translations.
Accents matter. Modern Greek uses the polytonic system only in formal or academic contexts with ancient texts. Contemporary Greek uses monotonic orthography — three accent marks maximum. A missing accent can change meaning entirely. «» (on/top) versus «» (which isn't a word but could confuse automated spellcheckers). Don't skip accent placement. It takes less than a minute to verify and saves headaches later. There are also calques to watch for. English loves phrasal verbs and prepositional constructions that don't translate cleanly. "Find out," "figure out," "carry out" — each one could mean three different things depending on context. Greek typically resolves these with a single, more specific verb. I always ask myself which Greek verb captures the intended meaning rather than chasing a literal rendering. The literal version almost never reads naturally. For learning resources, the Triandafyllides Grammar is the standard reference, though it's dense. For practical purposes, the online dictionary at perseus.tufts.edu handles ancient and modern forms, and the Greek-language news sites like kathimerini.gr give you a sense of how educated writers handle contemporary prose. Reading Greek journalism for thirty minutes a week does more for your translation instincts than any textbook exercise.
The tools themselves have limitations. DeepL's Greek support has improved substantially since I first encountered it around 2019. It handles technical vocabulary better than Google Translate and produces more natural word order. But it still struggles with idioms and tone. I've seen it render "break a leg" as a literal bone injury request, which is the kind of mistake that only shows up after you've already hit send on a project. Never trust a machine output without a full read-through. My current process: MT draft, then manual revision focusing on aspect and agreement first, then register and flow, then punctuation and orthography. That order matters because fixing aspect issues mid-pass often changes sentence structure, which then affects agreement and word order. Doing it in the wrong sequence means reworking the same section twice. If you're just starting out, pick a short text — a paragraph or two from a news article or product description — and translate it manually before touching a tool. You'll feel the friction between the two languages immediately, and that awareness saves you from making the same mistake three hundred times in a longer document.
