Research and Translation in Nepal

Nepal is not a one language country. Nepali is the language of government, schools and the national news, but the census records well over a hundred mother tongues. Outside Kathmandu, and particularly with older people and with women who left school early, Nepali is a second language spoken adequately rather than freely. An interview conducted in a contributor’s second language sounds acceptable on the day and lands flat in the edit.

This page is about that work: finding and approaching contributors, interpreting and translating across the languages people actually speak, working respectfully with communities and religious sites, and doing the desk research before you fly. It is not about general crew, which sits on its own page. We will not tell you that access is guaranteed. We can tell you who to approach and in what order, what they will want to know, what will make them refuse, and whether the subject is one where people talk at all.

Producers often meet the word fixer at the point where they need a permit chased or a driver booked. That is the visible end of the job. The part that decides whether a factual film works is the part that happens weeks or months earlier, when someone with local knowledge sits down with your treatment and asks the questions you cannot ask from abroad. Does this story exist in the place you think it does. Are the people you have imagined real, reachable and willing. Is the version of events in your synopsis the version people on the ground recognise, or a shape imported from a press cutting written by someone who never went.

Research in Nepal is not a database exercise. There is no comprehensive casting file for a farmer in Rukum, a returned migrant worker in Dhanusha or a nun in a Mustang gompa. You find those people the way local journalists find them, through the network of people who know a district: teachers, health workers, ward secretaries, the shopkeeper everyone walks past, the community groups that meet every month. That network takes time to work and it cannot be rushed by budget. A good researcher spends the first days listening rather than filming, and comes back with names, a sense of who carries authority in a place, and an honest read on whether your story survives contact with the people in it.

Fixing is the operational half of the same job. It is the phone calls, the ward-office visits, the negotiation over where a truck can park and who needs to be greeted first, and the quiet management of everybody’s expectations so that a crew of strangers can work in a place for a day without leaving damage behind them. Research finds the story. Fixing makes it possible to shoot. The two are usually done by the same people in prep and split into separate roles on the day, for reasons we come to below.

Casting for factual work here rarely starts with a database. It starts with people who know a district: local journalists, teachers, health workers, ward secretaries, community forest and mothers’ groups, the person who runs the shop everyone walks past. We build a shortlist from that, cross-check it against whatever record exists, and then approach in the order that causes the contributor the least trouble.

Order matters more than foreign producers expect. In most places, speaking to the household and the ward before a crew arrives, rather than afterwards, is the difference between a welcome and a wall. Turning up unannounced with cameras reads as either officialdom or intrusion, and both close doors. A contributor who has been approached properly has had time to talk it over with a spouse, an employer or an elder, which means the yes you get is more likely to survive the drive to their door.

First contact is by phone, Viber or WhatsApp, usually several times, and often through a relative’s handset rather than the contributor’s own. Many older people do not carry a smartphone, and a son or daughter in Kathmandu or the Gulf is frequently the practical channel to a parent in a village. Email reaches almost nobody outside an office. Letters and formal introductions still carry weight with institutions, monasteries and local government, where a request made in writing and then followed up in person is taken more seriously than a call out of the blue. We match the channel to the person, and we do not over-promise. Nobody is told they will be on television at the first approach, because at that stage it may not be true, and a promise you cannot keep poisons the relationship you need.

There is more than one way to interpret an interview, and the choice changes both your rushes and your day. Consecutive interpreting, where the contributor answers and the interpreter renders it afterwards, gives you clean, usable answers and roughly doubles the time each question takes. It is the right mode for anything that matters editorially, because you get the contributor’s own voice on the track, uninterrupted, and a faithful rendering after it. Whispered simultaneous interpreting, where the interpreter murmurs a running translation to the director, keeps the room oriented and the questions flowing, but it contaminates the audio and encourages people to talk over one another. Decide which you want per set-up, in prep, not on the day.

Interpretation belongs on its own audio channel. Clean contributor audio is never compromised for it, which usually means the interpreter is off-mic or on a separate track that the edit can drop. A good interpreter is briefed on the editorial line beforehand so they do not summarise, answer on the contributor’s behalf, or quietly soften a question they find rude, all of which are natural instincts that wreck an interview. The interpreter needs to know that a long, rambling answer is often the good one and should be rendered in full rather than tidied into a sentence, and that a silence is the contributor thinking, not a cue to help.

There is a human dimension too. An interpreter is often the only person in the room the contributor trusts, and that trust is an asset if it is respected and a problem if the interview turns adversarial. On sensitive stories we tell the interpreter what the difficult questions will be, so they are not ambushed into softening or refusing them mid-interview. And where the day is long or the subject heavy, interpreting is genuinely tiring work that degrades without breaks. We plan for that rather than treating the interpreter as a fixture that runs all day.

Nepali will get you understood across most of the country, but understood is not the same as at ease, and factual film lives on people being at ease. The languages people actually think and grieve and joke in vary sharply by region, and matching your interpreter to the district rather than to the country is one of the highest-value decisions in the whole plan.

Terai districts along the southern plains run on Maithili, Bhojpuri, Awadhi, Bajjika and Tharu. The middle hills carry Tamang, Newar, Gurung, Magar and the Rai and Limbu languages, and the Rai languages in particular are a group of related but genuinely different tongues rather than one, so a Bantawa speaker and a Kulung speaker are not interchangeable. The high north speaks Tibetic languages: Sherpa in Solukhumbu, Lo-ke in upper Mustang, distinct varieties in Dolpo, Humla and the Tsum and Nubri valleys. The far west has Doteli and its neighbours. The Kathmandu Valley has Newar under the Nepali that everyone conducts business in.

Where a contributor’s Nepali is thin, the workable answer is a two-step relay: mother tongue to Nepali to English, with someone from that community doing the first leg. It is slower, it puts an extra person on the fee, and it needs setting up in prep because the right relay interpreter often lives in the district and does not advertise. It is also the difference between a person explaining themselves and a person answering questions. The flat, guarded, monosyllabic interview that so often comes back from remote shoots is usually not a shy contributor. It is a contributor working in their second or third language while a stranger holds a microphone, and it is almost always avoidable if the language is planned rather than assumed.

A signature on an English release, from someone who does not read English, is paperwork rather than consent. It may satisfy a form-filling instinct, but it will not protect the contributor, and increasingly it will not satisfy a broadcaster’s compliance team either. What works is an explanation, in the contributor’s own language, of the things that actually matter to the person in front of the camera.

That means telling them who is making the film, who will broadcast it, that it will sit online and can be watched again, that people in Nepal including their neighbours and their employer may see it, that they can stop at any point, and exactly what will and will not happen about money. It means checking that the explanation landed, by asking the contributor to say back what they understood, rather than reading a paragraph and pointing at a line. Where your compliance allows, we record that exchange on camera as well as taking the written form, so there is a record of informed consent and not just a signature.

Anonymity deserves particular honesty. If a contributor is being offered protection, we will say plainly whether it is achievable, because in a village of two hundred people a blurred face and a changed name is not anonymity. The neighbours know who has the disabled child, who came back from Malaysia with nothing, who fell out with the ward chair. A promise of anonymity that the place itself defeats is worse than no promise, because the contributor relaxes into a candour they would not have offered if they had understood the real exposure. We would rather a contributor declines than agrees on the strength of protection we cannot deliver.

Nepal is a country of active, working religious sites, not museum pieces, and the difference matters the moment a camera comes out. Many Hindu temples restrict non-Hindu entry to inner areas, and no paperwork changes that, because it is a matter of religious practice rather than permission. Permission to film inside a major religious site usually sits with that site’s own trust or management committee rather than with any government permit, so the film permit in your folder is not the document that gets you through the temple gate.

Some specifics recur. Cremations at the river ghats are filmed often and are still a family’s private grief, so you ask, and you accept a refusal without argument. Photographing the Kumari, the living goddess, should be treated as off limits unless you are told otherwise by the people who manage the household. Leather is unwelcome in some precincts, so belts, bags and camera straps can quietly cause offence. Feet are not pointed at people, shrines or food, and heads are not touched, which sounds trivial until a crew member steps over an offering or perches a case on an altar step. Removing shoes, covering shoulders, and asking before pointing a lens at someone at prayer are small courtesies that keep a location open.

A monastery answers to its abbot and its monastic committee, not to a ministry, and filming during a ritual is a conversation rather than a form. The festival calendar is its own force. Dashain and Tihar empty government offices and fill roads and flights for weeks, which is either your story or your obstruction depending on the film, and the day you want to shoot may fall on a day when nobody will discuss access at all. We brief crews on all of this before they land, because a protocol breach is not just rude. It ends the day, and sometimes the relationship that took weeks to build.

A good deal of what you need is on paper, in Nepali, in a district office, and it will not come up in an English-language search. Court files, land records, ward minutes, health-post and school registers, and local newspaper archives are largely unindexed in English and often not online at all. Retrieving them means someone physically going to the right office, knowing which one holds what, and being able to read and summarise the result. This is slow, unglamorous work, and it is frequently the difference between a film that stands up and a film built on a single foreign correspondent’s cutting.

Two things catch foreign researchers out, and both are about dates. Official documents are dated in Bikram Sambat, which runs about fifty-six years ahead of the Gregorian calendar, with months that do not align to it, so a paper dated Shrawan 2081 is not the year you would assume from the numbers. Convert carefully, because a slip of a year can misdate a land dispute, a court ruling or a death by a wide margin. Second, the government fiscal year runs from mid-July to mid-July, which is the period any budget, allocation or spending figure refers to. Reading a rural municipality’s budget as a calendar-year figure will quietly mislead your script.

Beyond the paperwork, research means checking claims against the record where a record exists, and telling you honestly when one does not. Some of the most sensitive stories rest on numbers that were never collected or on events that no office wrote down, and it is better to know that at the research stage than to build a sequence around a statistic that cannot be sourced. We tell you what is documented, what is testimony, and what is rumour, so you can decide how to treat each on screen.

The common failure is not a refusal. It is a yes that was never a yes. People here rarely decline directly, because a flat no to a visitor is itself a discourtesy, so a contributor who is uneasy will agree on the phone, agree again the day before, and then be at a wedding in the next valley when you arrive. Miss that soft signal in the research and you have built a shoot day, a drive and a crew’s time around someone who was only being polite. Reading the difference between a real yes and a courteous one is a local skill, and it is most of what you are buying.

The second common failure is the interview that technically took place: an hour of short, guarded answers in a second language, an interpreter compressing as they go, and nobody in the room able to hear that the good material went past untranslated. It looks like a successful shoot day. It reveals itself weeks later in the edit, when the transcripts come back and the sequence you were counting on is not there.

The rest of the cost lands later still. A contributor who never understood where the film would be shown sees it circulating locally and asks to be removed, sometimes after transmission. A family is delivered by an organisation with its own reasons for wanting that story told, and the story arrives pre-shaped to that agenda. Someone speaks candidly about caste, land, an employer or a local politician, the crew flies out, and that person still lives in that ward with the consequences. Where any of that leaves you legally is a question for your compliance people, not for us. What is certain is the editorial cost, and the real one to the contributor. Both are avoidable, and both are decided in the research long before anyone reaches for a camera.